October, 2019 Decentral Park Capital
The Current State of Telegram Open Network
A sleeping giant awakens
Author: Elias Simos
Privacy is not for sale, and human rights should not be compromised out of fear or greed. - Pavel Durov (April, 2018)
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Table of contents
Key Takeaways 3
About Decentral Park 4
Introduction 5 Crypto winter and a tale of valuations 6 TON: Special case or one of many? 9
Part 1: All things Grams 10 1.1. Tоken Economics 11 1.2. What is a Gram worth? 13 1.3. Finding liquidity 17 1.4. A TON of use cases 18 1.4.1. Demand-side use cases; payments, bots, content and Dapps 18 1.4.1.1. Monetizing bots and enabling payments between users, developers and advertisers 18 1.4.1.2. Telegram Passport 20 1.4.1.3. Dapps 21 1.4.2. Supply-side use cases 22
Part 2: The current state of the Telegram Open Network 24 2.1. Taking TON apart; a review of TON’s architecture 24 2.1.1 The TON Blockchain 26 2.1.1.1. The masterchain 26 2.1.1.2. The workchains 27 2.1.1.3. The shardchains 27 2.1.1.4. Instant Hypercube Message Routing 28 2.1.1.5. TVM (TON Virtual Machine) 28 2.1.1.6. Smart contracts 29 2.1.1.7. The role of validators, nominators, fishermen and collators 29 2.1.1.8. Programming on TON; Fift 31 2.2. Best of the rest; other key pieces of the TON stack 32 2.2.1. TON Payments 32 2.2.2. TON Storage 32 2.2.3. TON Networking 33 2.2.4. TON DNS 33 2.2.5. TON Proxy 33 2.2.6. TON Services 34 2.3. Developer communities & recent updates 34 2.3.1. TON core, TON Labs and the early settlers 34 2.3.2. Current state of development 35 2.3.3. Looking ahead 37
1
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Part 3: Discussion 38 3.1. All those in favour... 38 3.1.1. A huge market opportunity 38 3.1.2. A collection of promising technologies 39 3.1.3. A world class team on a mission 40 3.2. And those that stand undecided… 40 3.2.1. Many unknowns in the token economy 40 3.3. And all those against… 41 3.3.1. A vision too grand 41 3.3.2. An unstable environment for cryptoassets 44 3.3.3. A developer unfriendly experience 45 3.3.4. A series of competitors 47 3.3.5. A regulatory mountain to climb 48
Conclusion 50
Appendix 51
Important Disclosures 57
2
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Key Takeaways
If delivered as intended, with 250M (and growing) active users on its front-end, TON can be the gateway for cryptoassets and the related applications to “bank the unbanked”, and become the first ubiquitous discovery platform for Web 3.0 applications, akin to the App Store for Web 2.0.
However, at this stage, interoperability appears a long way out, as (i) it is common for blockchain networks to spend at least one or two years post Mainnet launch to achieve a stable state and (ii) other networks need to achieve maturity, in order for interoperability to make sense.
Over a 5-year horizon, once key pieces of the platform are in place and the token economy stabilizes, we expect the network’s value to surpass $20B, making Grams a top-10 protocol asset.
TON introduces a collection of highly promising technologies, but the unknowns in the feasibility of some of the parts, as well as the sum of the parts, are still many.
In the initial stages of the launch, TON appears to be less open to developers than its competitors. In a world where talent is scarce, we view that as a competitive disadvantage.
The instability in the current macro- and meso- environments, makes the launch phase of the Gram token economy, all the more challenging.
The unfriendly attitude of local governments in the countries where Telegram’s users are predominantly based, may dampen the adoption of Grams by users, by slowing the value flow from fiat to Grams.
With weak demand among crypto-native investors, few avenues for current Telegram users to acquire Grams, delays in delivery, the Foundations intent to not proceed with buybacks, and ~60% of the total supply of Grams coming to market in Year 1, we expect selling pressure to outweigh demand for the asset in the short-to-medium term.
3
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
About Decentral Park
Decentral Park is a family of funds that offers a diversified approach to digital asset and blockchain investing. We research, identify, evaluate and support teams, products and market opportunities within the blockchain world.
Decentral Park’s founder-led team brings multi-dimensional experience, expertise, capital and network and industry relationships to blockchain technology projects. We believe blockchain technology and cryptoassets have the potential to make a hugely positive impact in many areas of the global economy.
Decentral Park Capital’s Principals have a long track record of founding, funding, developing, managing, advising, scaling and exiting successful businesses in blockchain, consumer technology, consumer brands, real estate, brick and mortar businesses, venture capital, investment banking and private equity - with a specialty in emerging markets and emergent industries.
Decentral Park Capital II LP is a private investment fund (the “Fund”) organized for the purpose of investing in cryptocurrencies, decentralized application tokens, protocol tokens and non-fungible tokens (together, referred to as “Digital Assets”).
The Fund aims to capitalize on the opportunity presented by the current distressed valuations in the cryptoasset space. We believe there is significant alpha to be captured in backing projects with listed tokens, world-class teams, strong track record of shipping code and experience of a full market cycle behind them. We have developed a suite of tools and methodologies that both enable us to select the winners and empower us to maximize upside potential.
The Fund will be investing for the medium term (2 years) using asymmetric expertise and information advantage, fundamental analysis and proprietary fundamental asset scoring and relative valuation tool (“DPC Heatmap”) developed by the General Partner to aid in the selection of undervalued tokens to maximize alpha return of its Digital Assets portfolio.
Cryptocurrencies as Digital Assets are an emerging, investable asset class that currently have very low penetration into both retail and institutional investors’ holdings. The General Partner expects high volatility in the short term, but medium- and long-term positive trends with considerable asset price appreciation as technological development and user engagement increases, and as investors increasingly view cryptocurrencies as an attractive asset class.
To qualify this report, the Decentral Park family of funds is not a holder of the TON SAFT.
4
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Introduction
The Telegram Open Netwоrk (TON) is a new blоckchain that will soon come online as a “TON is solving natural response tо the growing pains the blockchain ecosystem is currently subject to; namely (i) lоw capacity and transactiоn speed levels that pale in comparison to status for (i) lоw quo financial back-ends (e.g. the Visa network) and (ii) complex user interfaces (UI) and capacity and experiences (UX) оf cryptо prоducts (exchanges, wallets, digital identity management, transactiоn etc)1. speed levels and The project is the brainchild of Pavel and Nikolai Durov, Russia’s best known tech (ii) complex user 2 entrepreneurs, founders of social network VKontakte and the Telegram messaging app. interfaces in Over a decade, the founders of TON have created and scaled globally two era-defining internet ventures and are on the way to their third. What is particularly noteworthy is that production level they did so in the face of adversity, coming in direct conflict with the Russian government blockchains.” when it came to protecting the privacy of their users.
The ethos on which their businesses have been built and the personal brand they have developed along the way could not be a better match for the blockchain space. Telegram was born out of the founders’ need to communicate privately while under the magnifying glass of the Russian security services. While growing Telegram’s user base to the first 100M, the Durovs deployed a strategy of geographical arbitrage, frequently changing their base of operations between various cities in the world and building the messenger on a robust network of distributed servers. By merit of his track record, Pavel Durov - Telegram’s CEO, has become an icon among privacy activists in Russia and beyond. Telegram always looked more like a stateless cypherpunk messaging service than a glossy Silicon Valley type start-up. And TON appears to be here to carry that legacy forward.
Figure 1: From Telegram’s FAQ, on how they plan to monetize the application.
Among their many achievements, Pavel and Nikolai Durov have also spearheaded the 3 second largest fundraise in the history of blockchains , raising an estimated $1.7B over two private sales between February and March 2018; an impressive feat and inarguably
1 With Bitcоin ad Ethereum providing 7 and 15 transactiоns per secоnd respectively, the capacity оf TON blоckchain is expected tо be millions of transactiоns per secоnd - this is performance comparable with Visa and Mastercard. 2 ВКонта́ кте (vk.com) is the 15th most visited website in the world according to Alexa rankings (2019). 3 Second only to EOS - who raised $4B over the course of a year long ICO. The team was initially aiming for $1.2B in a part-private and part-public a’la ICO sale, but due to high institutional demand, decided to hold an only-private sale, and raise ~1.5x what they initially aimed for.
5
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
far more deserving than the majority of its neighbours at the top of the list of largest fundraises in the ICO era4. Now, a year and a half after TON closed its final $850M funding round and with the public launch expected in Q4 2019, the world looks like a fundamentally different place, both within crypto as well as more generally.
Crypto winter and a tale of valuations
“Since Feb 2018, February and March 2018 marked the onset of the second longest crypto winter, through the total market which, the total market capitalization of cryptoassets lost 85% of its value, top to bottom. Since a bottom was found at the end of 2018, the market has picked up, mostly fueled by capitalization of Bitcoin’s continuous ascent towards being adopted as a legitimate macro asset. At cryptoassets lost Decentral Park, we view the adoption dynamic in the blockchain world as bi-directional; 85% of its value, namely top-down and bottom-up. The top-down dynamic is driven by traditional institutions adopting cryptoassets as part of their standard rotation, while the bottom-up top to bottom.” dynamic involves grassroots innovation that leverages cryptoassets in order to massively improve on existing business and organization models (e.g. more markets via tokens, 24/7 liquidity, etc) and thrusting new ones into the limelight (e.g. programmatic markets powered by game theoretic primitives). Thus far, it appears that the top-down dynamic will play out faster than the bottom-up, in part because of the limited ability5 of Layer 1 blockchains to host large scale, high throughput applications.
Figure 2: Total crypto market cap has contracted by 1/3 since Feb/Mar 2018.
“Over 2017/2018 Further, the fact that cryptoassets are so new, comes in tandem with a lack of generally valuations in agreed upon valuation methodology. Due to that, over the course of 2017 and 2018 private market valuations in crypto were heavily skewed towards the “comparables” approach. crypto were However, unlike the public markets, once locked into said valuation, there is no formal heavily skewed way for private market valuations to adjust as the value of their comparables changes. towards the As a result, while publicly-traded cryptoassets lost a large proportion of their ascribed ‘comparables’ value over 2018, investments in the private markets were marked to book. At the same approach.” time, the 2018 bear market drove retail investors away, and the ones that are still around have sharpened their decision-making functions through their survivorship. The effect of
6
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
this change in landscape is that as these new assets break into the public markets, there “…as these new are fewer investors to pick them up from early investors’ hands and shepherd them for the remainder of their public life. assets break into the public markets, there are fewer investors to pick them up from early investors’ hands.” Figure 3: Top-10 ICOs by fundraise and subsequent returns until May 2019.
To highlight the effect of the “crypto winter” on valuations, over the past year and a half, the total crypto market cap lost ~25% of its value, while in the same period smart contracts platforms with liquid tokens have lost ~75% of their value. From the Top-10 ICOs of the 2017/2018 era, the ones that are liquid have lost 10% of their value on average, while the only one that remains profitable for its early investors is EOS. Removing EOS from the calculations paints a much more dire picture, with returns landing in the neighbourhood of a negative 70%. The intra-industry clusters that TON somewhat affiliates with have also performed poorly. On aggregate, equally weighted portfolios of smart contract platforms, Proof of Stake (PoS), Infrastructure and Interoperability related assets, have lost approximately 85% to 95% of their value since March 2018, signaling a lack of demand for the native tokens of these platforms.
“PoS, Infrastructure and Interoperability related assets, have lost
Figure 4: Aggregate performance of liquid tokens within the smart contracts, approximately infrastructure and PoS verticals. 85% to 95% of
The recent launches of Algorand’s and Hashgraph’s native protocol assets6 are indicative their value since of the effect that comparables-based valuations had in early stage blockchain investing March 2018.” over 2018/2019. Both projects are made up by teams of the highest caliber, backed by
4 Indicatively 3 out of the Top-10 raises have already gone to 0. 5 Albeit continuously improving. 6 ALGO and HBAR respectively.
7
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
some of the most recognizable names in venture capital, bringing to market technologies that - at least on paper - could revolutionize how blockchain technology is deployed at scale.
“It is a What the two projects also share, is the poor performance of their native tokens upon challenging time listing (ALGO is down 90%, while HBAR is down 80%). Indicatively, Algorand raised a total of $125M at a $7B supply-derived and a $24B implied valuation, while Hashgraph raised for assets in tech ~$100M at a $6B implied valuation. Both projects distributed less than 30% of their total transitioning treasury of tokens to their investors - introducing implicit overhang risk to the prospective from private to token holder’s decision model. Indicatively, the demand for these tokens is so low, that the daily OTC lending rate for HBAR tokens has been reported to be as high as 5% daily; public markets.” a hint to the fact that, at the moment, current holders are not interested in being long.
Figure 5: Performance of ALGO and HBAR after listing in Q3 2019.
“Investors are The disparity between private and public valuations is not a phenomenon exclusive to asking hard cryptocurrencies at this moment in time. It is no secret that a decade of quantitative easing and record low interest rates have been catalysts in driving capital towards venture capital, questions about as institutional allocators seek to take on more risk to meet their returns targets - which in the relationship turn has allowed thriving tech companies to stay private for longer. However, while venture between value investors are benchmarking growth over cash-flows, public market investors are not buying it. The result has been the lacklustre public market performance of highly creation and anticipated tech-IPO stocks. value capture.”
Figure 6: Popular tech IPO stock performance post listing in 20197.
7 Borrowed from The Information [Access here: http://bit.ly/2nvFVH7]
8
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
All of the above illustrate that it is a challenging time for assets in the technology sector transitioning from private to public markets. Investors and observers alike are asking
hard questions about the relationship between value creation and value capture in technology. Of course there is plenty of nuance in each of the examples here, and plenty to consider with regards to the potential upside that cryptoassets in particular still have. However, the evidence is fairly compelling that this is a landscape that requires investors to expect highly volatile listings and react and position appropriately.
TON: Special case or one of many?
Zooming back in to the main theme of this paper, TON’s case is hard to pigeon-hole. “TON is While it broadly falls into the category of assets described above, it really looks like none of the assets specifically discussed. Where blockchain platforms are struggling to attract coming online developers and users, TON is coming online piggybacking on a messenger app with piggybacking 8 approximately 250 million monthly active users (2x more than Uber, 4x more than on a messenger Zoom, 20x more than Slack, 500x more than Ethereum) and with the potential to become crypto’s definitive distribution and user acquisition platform9. This is the kind of potential app with that conjures up $1.7B in investment from some of the world’s premier patient capital, approximately reportedly including Benchmark, Sequoia Capital, Lightspeed Ventures and Kleiner 250M.” Perkins10.
Prior to the TON fundraise, the lights at Telegram were kept on by donations from the Durov brothers. Reports estimate the monthly burn rate to be approximately $1M, while the application is not generating any revenue yet. Claiming that TON is a broader effort to make Telegram a self-sustaining entity would not be a far reach - though it has the potential to become a lot more.
Overall, the Durovs’ pedigree, secretive mode of operations, and this market backdrop, “Reports make the upcoming release of TON’s Mainnet one of the most eagerly anticipated and attention grabbing Mainnet launches in recent years; one worth discussing, estimate the documenting and dissecting. In the following sections of the paper we will attempt to Messenger’s unveil much of the mystery around TON by diving deep into the recently released monthly burn documentation and code, aiming to (i) shed light into the utility of the Gram (GRM) rate at approx. tokens, by understanding how the Messenger and the Network will interface with one another and (ii) thoroughly understand the architecture technical composition of the $1M.” network.
8 Statista estimates Telegram’s MAU at 200M as of June 2019 [Access here: http://bit.ly/2ocHcUg]. Given that Telegram passed the 200M MAU milestone in March 2018, and its total user base stands at over 500M, we believe that 250M is a more accurate representation. 9 Provided that TON will be compatible and interoperable with all other major blockchains, something that this report is aiming to explore. 10 As reported by Coindesk [Access: http://bit.ly/2l6Lr2h].
9
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Part 1: All things Grams
“Grams will be This section explores the properties of the Gram11 token, both as the native utility token used both as a of TON and as an investment instrument. Telegram will be introducing Grams both as a consensus token that facilitates the process validators participate in order to reach consensus token consensus with respect to the state of the blockchain via Proof of Stake12, and as a money and as a money token to be used for anything between storing value, making payments on services token.” available via the messenger and on TON (e.g. Dapps), and transferring value between Telegram users (e.g. remittances).
In that respect, Grams can be assimilated to what Facebook is planning to launch through project Libra. The key difference is that while in the case of Libra the token will be pegged to a basket of currencies, the value of Grams will be - partly - decided by the market. We highlight partly here, as TON has a Central-Bank-like entity in place (the TON Reserve and
the TON Foundation) that can intervene to provide price support both in programmatic
ways, as well as with open market operations.
“The official launch of the TON Mainnet is scheduled to occur some time before the 31st of October 2019.”
Figure 7: Telegram users are the most interested in in-app money transfer features.
The official launch of the TON Mainnet and the official start of public sale of Grams are scheduled to occur some time before the 31st of October 2019. According to the Telegram Final Purchase Agreement the TON Foundation will have to return capital to investors should said deadline be missed. Provided that the network launches
11 Gram tokens are also notated as GRM tokens, the expected ticker the token will assume. 12 The PoS mechanism that TON is proposing on its whitepaper bears a lot of similarities to that of its interoperability competitor, Polkadot.
10
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
successfully by the 31st of October, Telegram will make Gram wallets available to approximately ⅔ of their user base13, gradually expanding to the whole.
1.1. Tоken Economics
The total supply of Gram tokens is initially set at 5 billion, with an inflation rate that will “The total be approximately 2% annualized according to estimates shared in the TON whitepaper. If that holds, the total supply of GRM will effectively double within 35 years from launch. supply of Gram As in most PoS chains, tokens generated via inflation will be earmarked as validator tokens is initially rewards. set at 5 billion,
We estimate that from the total supply, ~54% was sold to investors14 in at least 2-rounds with an inflation of fundraising. Of those, ~12% will be made immediately available for investors to trade. rate that will The remaining ~42% will be gradually made available, as it is subject to a varying lock- approximately up schedule that spans between 6 and 18 months post Mainnet launch. We further stand at 2% estimate that ~4% is earmarked as developer and community incentives, while the remainder15 will be controlled by the TON Foundation and TON Reserve. annualized.”
Figure 8: Expected GRM token distribution as of Mainnet launch.
While its final form is still unclear, the TON Reserve appears to be a minting facility16 that will be issuing and selling tokens according to a set pricing function. The prices will initially start at a low level, but will increase in an exponential fashion for every marginal Gram issued. The first 2.7B Grams to be issued are earmarked for SAFT holders.
Figure 9: The price function that TON has designated for the TON Reserve’s operation.
13 Estimated to stand at 300 million signed-up users - this information surfaced on the NYT and is attributed to unnamed TON investors. 14 Lock-up schedules (find) 15 Close to 50% of the total supply. 16 It is unclear whether the TON Reserve will be fully programmatic or a hybrid that involves human intervention.
11
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“The TON The fact that there will effectively be two markets for TON - a primary and a secondary - provides the issuer with one more lever to control the price of Grams by creating arbitrage Reserve opportunities, the parameters which are under their control. If the price of Grams is higher withholds the in the open market, more buying will take place via the Reserve and vice versa, eventually right to sell up bringing the primary and secondary markets at par with one another. According to the whitepaper, the TON Reserve withholds the right to sell up to 2.5B Grams (½ of total to 2.5B Grams.” supply) via the mechanism described above. Given that the estimated amount held by the TON Reserve at the moment stands at a little less than 35%, we can expect the remaining 15% to come from a mix of open market operations and validator rewards collected in the initial stages of the network’s launch17.
Figure 10: Jason Choi’s comment on the recent launches of ALGO and HBAR.
“TON is likely Arguably, the aforementioned parameters along with the patient capital that participated better prepared in the private sales and the massive war chest that TON raised in USD, mean that at the very least, TON is likely better prepared than other networks that have come to market in than other 2019 with regards to supporting the network token’s price in the initial stages of price networks that discovery. The above information also might imply that TON has the potential to become have come to one of the networks with more evenly distributed token economies in the years to follow. Given that (i) TON is planning to ultimately distribute 90% of the total token supply to market in 2019 stakeholders, and (ii) the supply side mechanisms in place to increase the probability of in supporting healthy price discovery in the long term, TON is poised to potentially be one of the the network networks with the lowest Gini coefficients18 in the blockchain space. token’s price.”
Figure 11: High level estimate of TON Gini coefficient at two levels of income (Top 100 addresses vs the rest) at launch and 18 months post-launch19.
17 It is still unclear whether TON will put the Reserve provision to work, as recent reports have it that were they to sell tokens directly to retail buyers, would risk the Gram being classified a security. 18 In economics, the Gini coefficient, sometimes called Gini index, or Gini ratio, is a measure of statistical dispersion that represents the wealth distribution of a population, and is the most commonly used measurement of inequality. 19 The Gini coefficients presented here, are a snapshot from February 2019. TON’s estimated Gini scores represent more illustrative assumptions, and less formally derived figures.
12
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
1.2. What is a Gram worth?
While the purpose of this paper is not to provide explicit price targets, the subject of price “Price is the one is inescapable. In fact, price is not and should not be a forbidden word or concept when thinking about nascent networks enabled and moderated by protocol tokens. Price is the variable to rule one variable to rule them all; the metric that contains the most information about an asset them all; the and the most singular and unifying indicator for all stakeholders in a token economy. In most singular the following section, we provide an overview of significant events that we consider as milestones in TON’s valuation and for the pricing of Grams with a view to arriving at a and unifying sober evaluation of the price action to ensue in the initial stages of the network’s launch, indicator for all and beyond. stakeholders in
a token economy.
Figure 12: Summary of significant price milestones for Grams.
The right to claim Gram tokens upon the launch of the network was sold to investors in “SAFT 1 opened two private sales that took place in February and March 2018. The first SAFT round opened in January 2018, at an estimated price of $0.38 per Gram. Investors participating in January 2018, at this stage are subjected to an 18-month vesting period. 81 investors took part, 3 of at an estimated them publicly disclosed their identities. During the second SAFT round in March 2018, price of $0.38 Grams were sold for $1.10-1.45, without a vesting period. 94 investors took part, whose per Gram; identities were not publicly disclosed. SAFT 2 closed in March 2018, at an average price of $1.33 per Gram.”
Figure 13: Estimation of Gram circulating supply over the first 36 months after the network’s launch20.
20 This represents the extreme case, where all investors make their holdings available to the market, as soon as their lock-up expires. It further assumes linear distribution of the Dev and Reserve funds over 36 months, 2% annual inflation, and the 10% TON Foundation allocations remaining unavailable.
13
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“Investors Each round was capped at $850M. Investors participating in these rounds were given allocations that ranged between $300k and $20M. According to TON, not more than 20% participating in of investments originated from Russian funds or individuals. For both rounds Telegram the SAFT filed a Form D 506c with the Securities and Exchange Commission (SEC); these offerings rounds were were made under a claim of federal exemption under Rule 506(c) and/or Regulation S given allocations under the Securities Act of 1933.
that ranged In March 2019, the little known Xena exchange21 listed a leveraged, cryptocurrency-settled between $300k derivative contract for the GRM token (XGRAM), which in the first month of trading reached and $20M.” up to $8 per XGRAM, though in a highly illiquid environment. At the moment, the price per XGRAM stands at $3.30, however, there is barely any activity in this particular market, with the last trade being recorded on the 4th of September 2019 and the first bid in the book standing at $2.22 - marking an effective bid/ask spread of ~50%. The Chinese exchange 22 Lbank has also listed a Gram/ETH futures pair, which at the time of writing trades at the equivalent of $1.00.
Figure 14: Gram futures (XGRAM/USD) at Xena exchange (as of 16/09/2019).
“Gram futures trade in highly illiquid markets.”
Figure 15: Gram futures (GRAMS/ETH) at Lbank exchange (as of 16/09/2019).
The latest more formal attempt towards public price discovery was made in July 2019, when Korean company Gram Asia, allegedly the largest holder of Grams in Asia, held a
21 This is a Russian exchange that launched in the summer of 2017. The only listed assets on the exchange are BTC, ETH and GRM futures (XGRAM). According to Forbes, they have raised approx. $3.5M in funding thus far. Xena Exchange was one of the early Telegram Passport adopters. 22 Lbank has been found to misreport traded volumes according to TIE. [Access: http://bit.ly/2lgwY3S]
14
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
23 physically settled futures sale for $4 per Gram. The sale took place on Liquid , an “Initial trading exchange entity affiliated with Gram Asia. Approximately 3.125M Grams were made available, raising $12.5M. The supply of future tokens in this case came from the in Grams can be notional allocation of Gram Asia, without though explicitly being given permission from expected to be TON. According to the TON private sale investment contract, the issuer does not allow volatile and the transfer of rights to future tokens before its Mainnet launches - however, depending on jurisdiction, there appear to be ways to override the clause. Finally, on the 19th of initial price September 2019, the little-known exchange ATAIX, launched a sale for allegedly discovery looks “unlocked” Grams - which in reality appear as unsecured futures agreements, for $3.99 likely to depend per Gram. on how In addition, multiple reports have surfaced hinting at a busy OTC market for Grams - aggressive though it appears skewed to the sellers side. According to a Coindesk interview with sellers are, and OTC trader Vladimir Cohen24, “OTC sellers have been striking confidential deals for Grams based on trust. Often, sellers are trying to resell their tokens for a profit, having whether - and at paid either $0.38 during per gram in the first round or $1.33 in the second.” In practical what price - a terms, what is being traded in the OTC markets are IOUs - effectively gentlemen’s stabilizing buyer agreements that the agreed upon amount of Grams will change hands, at the agreed upon price, once the Mainnet launches and the tokens are minted. More recently, we steps in to the have come across anecdotal reports that early investors are looking for liquidity OTC, market.” around the $1.50 mark, while some OTC brokers are reporting seeing offers at $0.85, with few bids.
Considering the highly volatile nature of the cryptocurrency market right now, and the poor early performance of other recent token issuances, initial trading of Grams can be expected to be volatile and initial price discovery looks likely to depend on how aggressive sellers are, and whether - and at what price - a stabilizing buyer steps in to the market. A repeated pattern is that once listed, token prices trend down towards the level of the earliest fundraising due to supply pressure from early investors .
At the same time, the equity research house ATON and cryptoasset investment bank Hash CIB have published two pricing models25 that follow a modified DCF approach. ATON’s price target lands at an average of $5 per Gram, while Hash CIB’s model sets a target at $5.90 for the end of 2019.
Once again, while not the explicit purpose of this paper, we cannot help but put the aforementioned targets to the test, based on our own set of assumptions. The methodology followed here is purposefully not following on the aforementioned models’ tracks, in order to add to the dimensionality of public knowledge. Our model
23 Liquid is the largest Japenese crypto exchange platform. More details on the sale can be found on the Techcrunch article that first broke the news. [Access: tcrn.ch/2l8ODKD] 24 [Access: bit.ly/2mauO5W] 25 ATON’s report benchmarks two prices, that represent their proposed fair value range - between $2.1 and $8. You can access ATON’s report on TON here [Access: http://bit.ly/2mcs7kj]
15
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
seeks to simply match an expectation of demand with the supply of Grams at any given point in time, assuming that all unlocked Gram tokens will be available for sale. While a simplistic approach, it establishes a supply/demand based support level for a Gram’s price.
“Our model The core assumption is that spending on Telegram functionality (via Grams) will start at seeks to simply ~35% of the current average user spend on the Apple App Store - an estimated $79 in 201826. Using that as the starting point and adding a few more assumptions27, we arrive at match an the following price targets. expectation of demand with the supply of Grams at any given point in
time.” Figure 16: DPC’s high level model parameters.
Figure 17: DPC’s demand projections for Grams.
Figure 18: DPC’s value bottom projections for TON.
One more underlying assumption to the model is that demand will hit the TON blockchain on day one. While somewhat unrealistic, given the early stage of the network’s development, we believe this assumption to be a useful proxy for estimating speculative
26 [Access: http://bit.ly/2kD5AwJ] 27 We assume that (i) TONs current MAU stands at 250M and (ii) will keep growing on the same (reported) linear trajectory since 2015 - discussed earlier in the report, (iii) the percentage of that spend funneled through Grams will follow said pattern, while (iv) the supply of Grams will follow the schedule outlined in the beginning of Section 1.2.
16
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
appetite in Grams. It further both benchmarks well on the most recent Gram futures “More recently, recorded prices and illustrates well the J-curve dynamic frequently observed in a cryptoasset’s life cycle28. Interestingly, our estimate of an approximate fair value bottom in a letter to is not dissimilar to the output of modelling work that preceded this paper, and is also in investors, the line with the implied provision for possible buybacks from the TON Reserve that are TON estimated to kick-in at the $1.80 price level29. More recently though, reports surfaced that in a letter to investors, the TON Foundation expressed their intent to not proceed with Foundation buybacks, in order to comply with SEC regulations and protect Grams from being expressed their classified as unregistered securities. Regardless, this approach can only work as long as intent to not the TON Reserve has enough capital to support it. proceed with buybacks, in 1.3. Finding liquidity order to comply with SEC
Token liquidity and price discovery depends on multiple and high quality exchange regulations.” listings. Given Telegram’s clout and TON’s considerable financial resources, this is not likely to be a problem. Coinbase recently published a list of tokens they are considering listing and the Gram is one of them. Further reports have surfaced naming the Binance exchange and the eTоrо platfоrm as prime candidates, while also hinting to a possible listing on Huоbi. However, thus far the only official announcement for a Gram listing has come from the Blackmoon Financial Group, an entity indirectly connected to TON’s executive team.
Blackmoon - a Caymans-registered exchange, with an estimated 3,800 user base - was founded in 2014 as a blockchain-based fintech company, while the Blackmoon exchange itself was launched in 2017 (product of a $30 million ICO) by Ilya Perekopsky (former vice-president of VKontakte), Oleg Seydak (co-founder of Flint capital)30. Among the advisors, board members and investors of Blackmoon are Vasyl Latsanych (Vice President of Major Markets at Veon), Oskar Hartmann (KupiVip.ru, Carprice.ru, board member of Alfa-Bank), and Sasha Ivanov (founder of Waves). The supply of Grams for the Blackmoon listing will allegedly come from the Gram Vault, one of the few custody providers for Grams - thus likely becoming an intermediary in facilitating early investors’ path to liquidity.
28 Chris Burniske’s thesis on the crypto J-curve, foresees a boom and bust cycle early in a cryptoasset’s life cycle, as Discounted Expected Utility Value is substituted for Current Utility Value. [Access: http://bit.ly/2l9h0s6] 29 According to the whitepaper, after the allocations to the TON Foundation and the founders, the minimum price for Gram should be effectively set at $1.81. TON Reserveis poised to engage in buybacks if the price drops below 50% of that calculated by Reserve’s Gram pricing formula. 30 According to Techcrunch [Access here: https://tcrn.ch/2mkVst3]
17
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
1.4. A TON of use cases
As we briefly touched upon in previous sections, Grams will have 2 primary streams of functionality; as a (i) demand side payment token and (ii) as a supply side consensus token31. In this section we will unbundle the two streams of use cases and examine exactly how they will find utility within the current and projected future Telegram ecosystem.
Figure 19: The Telegram Open Network stack as proposed by the whitepaper.
1.4.1. Demand-side use cases; payments, bots, content and Dapps
1.4.1.1. Monetizing bots and enabling payments between users, developers and advertisers
At the mоment, we estimate that there are mоre than 800,000 Telegram-Bоts, which are used actively by approximately ¼ of the whole Telegram user base. Telegram Bots are effectively accounts operated by AI software, that enable a series of actions on the platform, such as semi-automated payment processing, group management, game notifications, donations and more, by reading and sending messages, commands and requests to and from users. These messages and interactions are funneled from the Telegram API to the software running on the servers of each particular bot developer via a HTTPS-interface.
One of the few applications that have already integrated with the TON Testnet is the ButtonWallet - a multi- currency wallet and P2P transactions service hosted wholly within the Telegram Messenger and functioning as a bot, rather than an app.
31 Demand-side use cases for the Gram tokens modules higher up the technology stack that makes up TON, while supply-side use cases are found further down the stack.
18
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“We expect the TON Payments platform to not be ready for prime-time, by Mainnet launch.”
Figure 20: Button Wallet user interface and functionality.
The Telegram payment platfоrm is currently integrated with 8 payment processors32 that allоw bоts tо accept credit card payments frоm users in over 200 cоuntries. By introducing Grams into the ecosystem, Telegram will provide the users, developers and advertisers on the platform with an alternative avenue to fulfilling payments on the platform; one that will presumably be a lot more competitive than the 2.9% and $0.30 that Stripe - for example - currently charges per transaction.
The launch version of TON will come with a Wallet integration in the Messenger. At this stage, however, we expect the TON Payments platform to not be ready for prime-time33 and as such, transactions will take place exclusively on-chain, implying higher transaction
fees and less throughput.
Further, there is a point to be made about volatility and its taxing overhead - both for “There is a point consumers, as well as developers and businesses running on Telegram. Thus far in the development of cryptoassets, we have no tangible demonstration of a volatile cryptoasset to be made being used widely as means of payment. Perhaps the most successful asset in that regard about volatility is ETH (Ethereum’s native token) and its utility as payment token for executing work for and its taxing the network34 - but it is still far from ubiquitous. overhead.” At the same time, stablecoins are quickly getting traction, currently standing at over $5B in market capitalization and trading at volumes in the multiple of at least 4x their market capitalization. Nothing stops TON from launching their own stablecoin, although at the time of writing, there are no confirmed plans for such a feature.
32 These are: Stripe, Paymentwall, Yandex.Money, Sberbank, Payme, CLICK, Rave by Flutterwave and Tranzzo [Access here: http://bit.ly/2mtnS3R] 33 TON Payments” is a micropayment channel network - similart to the Lightning network for Bitcoin, aiming to offer instant off-chain value transfers between users, bots and other services, at a level of security equivalent to that of on- chain transactions. 34 via Gitcoin, in and around ETH Global events etc
19
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
In the above context, an interesting layer in the stack that is at present likely missing from “While not the available information space, is a link between users that want to pay in Grams, and directly related merchants that want to be paid in fiat currencies. This is an interesting space for market to the first-order markers to operate in, although to the best of our knowledge, Telegram has not (yet) utility of Grams, provisioned for this type of liquidity pool.
the Telegram 1.4.1.2. Telegram Passport Passport is an important While not directly related to the first-order utility of Grams, the Telegram Passport is an important building block in enabling payments on TON. With critical user data stored in building block the Telegram cloud, encrypted end-to-end, one can imagine massive improvements in in enabling user experience from something simple such as buying airplane tickets through a payments on Telegram bot, to signing up to a lending facility and instantly going through the otherwise laborious KYC/AML process in order to apply for a loan. While signing up for the Telegram TON.” Passport will not be required in order to send or receive Grams, it will be obligatory for users interested in interfacing with an exchange venue in order to purchase Grams, via the Messenger.
The Telegram Passport is currently available in a test version via the Messenger with step- by-step instructions available for users to navigate its set-up35. The functionality of the Passport can be summarized in 3 steps:
The user locally encrypts his/her personal data (name, email, passport scan, other documents) with a password. The encrypted data + meta-information are uploaded to the Telegram cloud. When the user needs to log in to the service, the client downloads data from the
cloud, decrypts it with a password, re-encrypts it to the public key of the service
that requested the information packet.
“In the initial In the initial stages of the launch, user information will be stored in Telegram servers, until stages of the TON releases a decentralized storage component, as foreseen by the whitepaper. It is worth noting that this decision - especially given that third-party, pluggable, decentralized launch, user storage solutions are available in the market - leaves a potentially critical window for information will exploits to take place, that could be detrimental to Telegram’s reputation and the trust it be stored in has built with its user base, as it is readying TON for prime time. While Telegram’s reputation has been built on being an ultra secure way to communicate with peers, there Telegram have been occasions where these attributes have been put to test. The latest one is an servers.” incident where the identities of, at least, 85 protesters in Hong Kong were leaked to the Chinese government due to a security vulnerability.
35 Telegram Passport Manual - https://core.telegram.org/passport
20
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“Telegram’s Passport employs a SHA- 512 hash generator for the encryption - one known to be fairly simple for a sophisticated attacker to
break.” Figure 21: The Chinese government reportedly uncovered a vulnerability in Telegram’s security and was able to extract the identities of 85 protesters during the recent upheaval in Hong Kong36.
Further adding to the concerns about its security profile, the Passport references Telegram’s security protocol MTProto 2.0 directly via API, which employs a SHA-512 hash generator for the encryption - one known to be fairly simple for a sophisticated attacker to break37. Despite the fact that there are several password hashing functions that make brute forcing too complex for attackers (e.g bcrypt, Argon2, scrypt, PBKDF2), TON developers chose to not use these in the latest iteration of the security protocol. Simply put, the encryption user data stored in the Passport is critically dependent on the complexity of the password chosen by the end-user - and thus introduces a weak point in the system. Telegram has always opted for using their own proprietary encryption algorithm, breaking a common rule among cryptographers to never use one’s own cryptography38. While the more recent update that they pushed in 2018 (MTProto 2.0) was received as a massive improvement over the previous iteration39, the community is still puzzled as to why the developers would not simply choose a suite of far more extensively peer reviewed algorithms like "SPHINX" or "Pythia".
1.4.1.3. Dapps
One wholly new stream of functionality that will be gradually introduced as TON matures, is a universe of Dapps available to Telegram users40. Users will be able tо access that
36 Telegram has since explicitly addressed the issue [Access here: https://zd.net/2mtBw76] 37 10 GPU will test all combinations of 8-digit passwords from a 94х symbol dictionary (english letters, numbers, special symbols) in less than 5 days. 38 Also known as “don’t roll your own crypto”. 39 Some of the flaws of MT Proto 1.0 are explored in depth by Jakobsen and Orlandi (2015). [Access here: http://bit.ly/2n6MgsA] 40 Shorthand for decentralized application.
21
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“Backwards through an App Store-like search utility. Applications will be sоrted by pоpularity ands user will receive recommendations based оn their search histоry and downloaded Dapps. compatibility In-app premium features are expected to be paid for with Grams. with Ethereum, will be enabled Further, the TON blockchain will be - in theory - interoperable with Dapps built on Ethereum. As such, Ethereum Dapps that demonstrate product/market and by a Solidity product/platform fit with TON and the Telegram Messenger ecosystem, could be compiler built integrated with TON. Backwards compatibility will be enabled by a Solidity compiler built by TON Labs.” by TON Labs, a startup helmed by lead investors in Telegram’s token sale. The compiler has been in development since July 2019 and will be available for testing after the Mainnet launch.
While it is too early to make assumptions about which Dapps (if any) might be ported over to TON - or at least made interoperable - it is worth noting that the architecture of the TON blockchain is robust enough to enable even the Dapps with the highest throughput requirements. Whether or not Ethereum Dapps will choose to migrate over to TON will depend on various factors, including the security, integrity and robustness of the TON blockchain, the performance it enables and the fee structure it exposes developers and their users to. And besides, interoperability on TON, is a feature that we are not expecting
to see until after 2021.
“Interoperabili- There are many elements in the equation that are still unknowns. However, in an ty on TON is a interoperable world, with TON introducing hundreds of millions of users to cryptoassets and Ethereum’s state growing increasingly more valuable, one can imagine the two feature we are working together as complements; TON referencing Ethereum’s state (e.g. to give access not expecting to to a decentralized lending facility to link Maker DAO to their users, via the Telegram see until after messenger) and Ethereum leveraging the Telegram messenger as the discovery platform and superior user interface that it so desperately needs. 2021.”
1.4.2. Supply-side use cases
The past 10 years in the blockchain world have produced a series of replicable supply- side models (e.g. consensus algorithms) that more-or-less work, depending on the use case. As such, it takes less logical jumps to arrive to use cases for Grams when searching for utility further down the technology stack. The TON whitepaper makes the distinction between TON Services and Applications, with the former relating to the network layer of TON and the latter referring to the user facing end. The Services end will be seeded with functionality41 from the TON Foundation, starting with Storage, Proxy and DNS42, however, the expectation is for an ecosystem of validators, protocol developers and other parties to participate in the network-side economy. Below we provide a list of the use cases we see Grams being deployed in, in the supply side of TON.
41 This is a provision seen in the whitepaper, the state of development of which, is still unknown. 42 We explore these in depth in Part 2.
22
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
a. Staking from validators and nominators in order to participate in TON’s variant of “It appears that PoS consensus. b. Payments tо nоdes/validators fоr verifying transactiоns and smart cоntracts43. TON is c. Payments to the masterchain in order to publish a new workchain. learning from d. Vote on governance decisions (e.g. regarding the adoption of changes on the the successes TON protocol). e. Payments fоr stоring files in TON Stоrage - distributed file stоrage system with and failures of access tо Torrent technology and the use оf smart cоntracts tо maintain stability. those that came f. Payments for TON DNS and WWW (Wоrld Wide Web) - similar tо a DNS system before them, by that prоvides synchrоnizatiоn оf IP addresses and dоmains. The TON DNS will prоvide the ability tо have "human" names fоr accоunts, smart cоntracts, bridging the services, and nоdes. supply and demand sides of Overall, what TON proposes is a fully-fledged ecosystem in which Grams find ample use cases. The architecture of TON, which we will explore in Part 2, is an amalgamation of the network, various propositions that have been put into production from TON’s proponents in the through the use blockchain space. The type of utility that Grams are poised to find, follow suit; it appears of Grams.” that TON is learning from the successes and failures of those that came before them, by bridging the supply and demand sides of the network, through the use of Grams.
We often see blockchain based networks - especially those jumpstarted by incumbents with established fiat revenue streams - concentrate the token flow within the supply side of the network44. We think that design decision impairs the long term viability of a token economy, as effectively it leaves only one value absorption gateway open - from Dapp developers to validators45, and splits the ecosystem into two, limiting the potential network effect the token economy can achieve. TON introduces enough value sinks in the token flow, that - in theory - should provide a good foundation for the Gram economy to grow on. However, given some of the demand-side risks utility tokens face, it also appears that ensuring that end-users are not wiped out by volatility is table stakes.
43 TON will have around 100 validator nodes confirming each block on the chain. Each workchain will have 1024 smaller validators who are selected pseudo-randomly every 1024 blocks. 44 Refer to the Appendix (Figure 1) for a schematic of the differennt types of value flow in modern day blockchains. 45 e.g. developers collect fiat from users and with some of that, they buy and stake tokens in order to secure “real estate” on the network and pay validators for network resources.
23
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Part 2: The current state of the Telegram Open Network
“TON positions This part оf the repоrt is dedicated to abstracting away the complexity in the architecture itself as an entire of TON, by examining its components, as proposed by the documents TON has made public, and adding to the context around those with an evaluation of recent developments ecоsystem, in the ecosystem. presented as an amalgamation of Part 2 was primarily based on a review of the: TON Whitepaper 46 some of the TON Blockchain description47 most dominant Telegram Open Network Virtual Machine description 48 ideas that have TON Labs Toolchain and related documentation49 50 surfaced in the TON SDK Client Library and related documentation
blockchain space over the past 10 years.”
Figure 22: An overview of the defining characteristics of preceding and competing protocols as proposed in the TON whitepaper.
2.1. Taking TON apart; a review of TON’s architecture
TON positions itself as an entire ecоsystem, and is presented as an amalgamation of some of the most dominant ideas that have surfaced in the distributed web and cryptoasset space over the past 10 years. While there appears to be less innovation in each piece that makes up the whole, there is a lot in the assemblage of it all. The main aim of TON is tо sоlve for the scalability problem in distributed networks of computation and value transfer,
46 [Access here: https://test.ton.org/ton.pdf] 47 [Access here: https://test.ton.org/tblkch.pdf] 48 [Access here: https://test.ton.org/tvm.pdf] 49 [Access here: https://github.com/tonlabs/ton-client-rs] 50 [Access here: https://docs.ton.dev/86757ecb2/p/09bb3d]
24
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
with the ultimate goal of facilitating millions of transactions per second. TON aims to “TON solves achieve that by sharding and Proof of Stake51. Sо far, there are several PoS blockchains functioning but there are no production ready, shard-based blockchains - though there for scalability are several in development that are tightly or loosely endorsing the approach, including and Ethereum 2.0, Polkadot and Cosmos. interoperability
TON is, in effect, a net of blockchains, separated into 3 distinct layers: by designing The masterchain; a settlement layer that binds the system together - similarly to according to the 52 the Beacon chain in Ethereum 2.0’s or the Cosmos Hub in Cosmos’ designs . Infinite The workchains; blockchains with their own set of rules that reference back to the masterchain to sync on global state of the TON ecosystem. Sharding The shardchains; use-case specific blockchains that reference back to the Paradigm.” workchains for their specific strand of functionality. This is where user level interactions take place in the TON ecosystem.
Figure 23: High-level schematic of TON’s architecture.
TON’s architecture is dynamic in character. TON solves for scalability and interoperability by designing according to the Infinite Sharding Paradigm53, with a bottom-up approach, where each shardchain is perceived as its own account. These accounts are then grouped together based on their proximity in the network space54, allowing for messages
51 While sharding is a common concept in (centralized) database engineering, since it is unproven in a blockchain paradigm, it often attracts a lot of criticism. For those interested, we have prepared a set of notes to help a deeper dive into sharding. [Access here: http://bit.ly/2lzhisQ] 52 We encourage interested readers to review our synopsis of the Ethereum 2.0 roadmap for context. [Access herre: http://bit.ly/2mFxnx6] 53 The premise here is that data is represented in such a way that divisibility of data structures in TON is possible at a granular level. The whitepaper presents this idea, as “everything is a bag of cells”, where data is represented as cells with up to four pointers to other cells, which makes cells interconnected at the lowest level. 54 Benchmarked by the numerical proximity of their 256-bit addresses.
25
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“In order to be to travel across the network at high speeds. In order to be resource-efficient, shards that undergo activity overload can be split, while the opposite is true for shards with low activity resource allowing for messages55 to travel across the network at high speeds. In order to be efficient, shards resource-efficient, shards that undergo activity overload can be split, while the opposite 56 that undergo is true for shards with low activity - all while maintaining continuity of state . Block time is set to approximately 5 seconds, which will expectedly lead to short transaction activity overload confirmation times57. TON insulates the system against hostile forks by slashing the can be split and apostate validators’ stakes. While as in all PoS systems it is still theoretically possible to merged back mount a Sybil attack, presumably by the time the state on TON becomes so valuable so as to entice the majority of validators to collude, their stakes will be worth more than the again.” expected value of the fork58.
2.1.1 The TON Blockchain
This is the beating heart of TON. The blockchain is responsible for all value operations, storing and executing smart contracts, as well as interoperating with other blockchain ecosystems.
2.1.1.1. The masterchain
“The The masterchain is a specific chain (with workchain_id = -1), and acts as the base layer masterchain is for the network and the connection dock between all other shardchains. The masterchain is responsible for governance (validator list) and for base layer consensus in TON’s PоS. respоnsible for It is the layer that guarantees security for the rest of the network and contains information gоvernance and about the status of all wоrkchains and shards. This is the layer also responsible for hosting base-layer smart-cоntract state. The smart contracts in this layer might manage anything from the validatоr list to the inflation for the Gram token economy. The set of validators that secure consensus in the masterchain will start at 100 and might increase to 100059. The hash of the last TON’s PоS.” masterchain block determines the overall state of the system, such that monitoring the state of all shardchains separately is not necessary. TON is currently using SHA-256 as the hashing function. If this turns out to be weaker than expected, it can be replaced by another hash function in the future. The choice of the hash function is a configurable parameter of the protocol, and can be changed without a hard fork, through voting.
55 Messages are the primary vehicle for information exchange on the TON Blockchain. That includes any type of value transfer or state change update. Messages incur fees, as every other operation on the chain. 56 The whitepaper provides an example of overload and underutilization conditions; for the former it benchmarks 90% of the block being full for 5 or more consecutive minutes, while for the latter it benchmarks 60% of a block’s computation limit (size) for the two “sibling” shardchains combined. 57 The chosen interval appears unrealistically fast - considering the time at which the first iteration of the whitepaper was published (December 2017), in a social context, claiming ultra-fast transaction capability was table stakes. 58 In Proof of Stake the risk of forking or malicious attacks is never perfectly 0 - i.e. a highly motivated, cost agnostic actor, could - in theory - break the system. In practice this is highly unlikely. 59 There are still many looming questions about (i) how validators will be selected and (ii) how consensus and governance will be streamlined as the list of validators grows towards 1000.
26
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
2.1.1.2. The workchains
The workchains are separate, domain specific blockchains that run on limited “To become functionality virtual machines, and are bound by their own rules, programming language active a new and cryptoeconomic primitives. Workchains are similar to parachains in Polkadot’s, or zones in Comos’s design. At launch, TON will come only with workchain zero workchain must (workchain_id = 0) enabled - the chain that contains the initial state and logic for Grams have ⅔ and TON Smart Contracts. It is expected that most applications will only require consensus from workchain zero. validators.” Anyone that is part of the network can create a new workchain if they pay a masterchain transaction fee, required to publish the formal spec for a new workchain. To become active, a new workchain must have ⅔ consensus from active validators. Validators will have to upgrade their software to handle blocks/transactions on this new chain, so the parties behind the new workchain may need to offer some sort of incentive. Workchains produce blocks with 1-second delay to other shardchains and store the root hashes of the states of other chains. TON provisions for 2³² workchains, where each workchain can host its own native cryptoasset (currency_id = n), effectively raising the number of
of cryptocurrencies that can be spawned on TON, to 2³². As with workchain zero, Grams
are annotated as currency_id = 0, while other cryptoassets that will live on TON, will
sequentially assume currency_id’s that span from 1 to 2³².
Workchains play an integral role in enabling interoperability between different chains as “Workchains their design allows for multiple programming languages to be deployed on top of them. allow for To illustrate, a developer could implement the light client node logic of Bitcoin (with code written in C), with a custom coin (such as wrapped BTC) and references to the Bitcoin multiple blockchain, in order to link TON with the Bitcoin blockchain. In other words, the programming developer can write a smart contract that will read data from a Bitcoin node, check that languages to be the balance is updated on a specific address, and then mint wrapped BTC tokens in the TON workchain60. deployed on top of them.”
2.1.1.3. The shardchains
Each workchain is in turn subdivided into up to 260 shardchains that are bound to the same rules and block format as the workchain they are attached to, but are responsible only for a subset of accounts. Each shardchain is allocated a random set of validators
60 This is potentially one more interesting space for market makers to operate in, arbitraging on the latency between message propagation between blockchains.
27
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
61 62 “…this particular responsible for creating and validating new blocks . Each block within a shardchain is conceptualized as its own blockchain63, so that if invalid blocks need to be changed, a feature new block is committed that contains the residual information necessary to correct safeguards whatever imbalance the invalid block caused64. This particular feature safeguards TON TON from from chain rollbacks that can cause a lot more damage in the process of restoring state to a previous timestamp, by recalling all valid transactions that took place between the chain rollbacks, event and the initiation of the rollback65. that can cause a lot of damage in 2.1.1.4. Instant Hypercube Message Routing the process of restoring state Interoperability between shardchains is facilitated by Instant Hypercube Message Routing to a previous (IHMR) - a protocol for propagating messages across shards. Each shard is connected to other shards differing in exactly one hexadecimal digit of their shard identifiers. All timestamp.” shardchains constitute a “hypercube” graph66, and messages travel along the edges of this hypercube. In IHMR, messages travel along a slow path or a fast path. On the slow path, messages hop from one neighboring shard to the other - a process that guarantees delivery albeit being slow and expensive, while on the fast path, messages can travel directly between shards that lie further from one another in the network space67. The two approaches run in parallel. If a proof of the success of the fast method is committed to an intermediate shardchain, the slow method can be aborted. Either way the message gets delivered. The array of sent messages between shards is a complex multidimensional DAG. IHMR effectively enables the delivery of a message created in a block of one shardchain into the very next block of the destination shardchain, regardless of the total number of shardchains in the system.
2.1.1.5. TVM (TON Virtual Machine)
Virtual Machines are the computation layer for smart contracts in the blockchain world. The TVM is used to execute smart contract logic in the masterchain and the base workchain, can run several processes simultaneously and packs data tight68, making it
61 The exact process with which validators are selected is unclear. 62 Shards have twо parts; a split and a nоn-split. The split part cоnsists of accоunt, smart-cоntract, balance and transactiоn histоry information. The nоn-split part cоnsists оf the infоrmatiоn abоut all the accounts, with which this shard is cоnnected. At the moment, it is unclear in which cases of cross-shard communication only the header of a block or the full body of a block will be transmitted. Without clear definition this could become a point of a double-spend attack on cross-shard transactions. 63 Referencing the Infinite Sharding Paradigm. 64 In effect, the best visual proxy for TON’s architecture is that of a snowflake. [Access here: http://bit.ly/2n5M50v] 65 It is as of yet unclear what the resource prioritization will be in shards that need repair (i.e. smaller first, larger first or global repair?) 66 Refer to Figure 3 in the Appendix for an illustration of a hypercube graph. 67 Both mechanisms are deployed in parallel, and the slow routing is aborted only when a proof that the fast path was successful in delivering the message. The slow path guarantees delivery and is therefore used to ensure that crucial information about state updates is never missed (on the fast path, it is more likely that the validators of the destination shardchain, might “miss” the broadcasted message). 68 Data is represented as a collection of cells. Each cell contains up to 128 data bytes and 4 references to other cells.
28
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
efficient and fast. This also implies that its design supports interchain compatibility, such “The that validators from any chain can verify transactions on any other chain. The architecture of the TVM makes it more suitable to execute code written in higher level languages like architecture of Java - however, these will not be supported out of the box. The benefit here is that high the TVM can level languages are easier to code, debug and maintain. The flipside is that programs host higher level written in high level languages are comparatively slower, less memory efficient and cannot communicate directly with the hardware (such as servers etc). The memory languages like efficiency aspect can become a thorny issue down the line, as deployment on TON Java - however, appears to be expensive. Whether it will prove to be a feature - by forcing developers to these will not be prioritize-efficiency with the programs they deploy, or a bug - by deterring experimentation and thus to some extent iterative development, remains to be seen as supported out adoption ensues. of the box.”
2.1.1.6. Smart contracts
TON makes a distinction between local and global smart contracts; the former need not be aware of the global state (reference the masterchain), while the latter do. The way
smart contract logic works in TON is almost like a hybrid between Bitcоin and Ethereum; the message system is similar to the one that Bitcоin employs for state updates, while the functionality оf smart-cоntract cоde, addresses, etc. are similar to Ethereum. The disadvantage here is the price of operation. To illustrate, when a smart contract is
deployed on the TON blоckchain, there is a requirement for paying rent (in Grams), even
if the contract is not in use - which introduces an overhead for developers, in that they
have to use their allocated storage wisely.
This idea is contrary to Ethereum’s architecture, where the user pays (in Gas) just for the “…the advantage 69 amount of data they consume, but not for the time . On the flipside though, the of such a system advantage of such a system is that it forces developers to be careful with what they deploy on-chain (removing clutter) and introduces an additional layer of security. TON is that it forces also promises to deploy parallel sending and writing of smart-contracts. This feature, developers to be known as multithreading, allows working simultaneously with high load project with careful with multiple computations on different smart contracts. At the time of writing, it seems that this is a feature to be deployed down the line. what they deploy on- chain.” 2.1.1.7. The role of validators, nominators, fishermen and collators
As in all PoS blockchains, validators70 are nodes that are tasked with confirming and reporting the state of the blockchain truthfully to the rest of the network. A validator operation is a non-trivial activity that requires considerable disk space, computing
69 This approach is not efficient, as it introduces additional clutter on the Ethereum chain and will be replaced in future versions of the Ethereum platform. 70 Validators in PoS, equate to miners in Proof of Work (e.g. in Bitcoin)
29
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
power, and network bandwidth. Validators on TON will use an efficient variant of the BFT “In order to (Byzantine Fault Tolerant) consensus protocol, similar to PBFT or Honey Badger BFT. become a Each validator may be in multiple validator subsets and be expected to run validator validation/consensus algorithms in parallel. In order to become a validator candidate, interested parties will have to stake a minimum of 100k Grams. In order to become an candidate, active validator71, however, it is likely that the minimum requirement of Grams will be much interested higher72. Active validators perform useful work for the network and in return, receive parties will have inflationary rewards (in Grams), proportional to their stake. to stake a Validators need to run 2 smart contracts; the controller & the elector. The former interacts minimum of with other contracts to run the validator, while the latter enables one to participate in 100k Grams.” validator elections or collect unfrozen stakes and bonuses. Validators must ensure that they keep uptime, perform computations requested by smart contracts, receive updates about other shardchains etc, or risk getting slashed73 partly or fully, and even getting permanently excluded from the validator candidate pool. The estimated token denominated annualized return for TON validators stands at approx. 20% of the validators 74 stake - according to the whitepaper . If this is indeed true, it would place Grams among the highest yielding PoS cryptoassets in the Top-100.
Figure 24: Annualized (token denominated expected) returns in popular PoS networks.
“The estimated The global set of validators is known about one month in advance75, while the local set is established about one hour in advance76. The validators’ stakes are frozen for two months annualized in order to make sure that validators are slashed if they are proven to misreport on the return stands at blockchain’s state by fishermen or other network participants. The algorithm that selects approx. 20% of task group validators each hour, is pseudo-random. All of the parameters for choosing the validators validators could be later changed though voting. After the proposal is put forward, 2/3 of the total number of network validators need to agree to it, in order for it to take effect. stake.” There are also provisions in place, in order to limit the power of validators with large Gram holdings – they are however, still unclear.
71 i.e. actually performing work and receiving rewards for it. Not all validator candidates will be active validators. 72 Which could be as high as 10M Grams, according to reports from the early validator community. 73 Slashing refers to the validators stake being burned. 74 This could translate to 2% newly minted and 18% for tx/proc fees. 75 Or 219 masterchain blocks. 76 Or 210 masterchain blocks.
30
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
In rounding out the entities that help the system reach consensus, TON appears to have “Nominators borrowed a page from Polkadot77, introducing nominators, fishermen and collators, as consensus making parties. Below, we give a brief description of the role each of these piggyback on parties fill in TON. validators’ infrastructure Nominators are “delegators” that lend their holdings to validators, with a view to profit sharing. Effectively Nominators piggyback on validators’ infrastructure and and large initial large initial stake in exchange for a % off the proportional rewards. The return stake in profile that nominators are exposed to depends wholly on the validator’s exchange for a % performance. Present validators, nominators "vote" for validators by lending them Grams. Nominators can be individual holders of Grams or staking pools off the validator who will collect Grams from TON users and send them to the TON smart contract rewards.” to become a validator, acting like delegates. If they succeed, the members of the staking pool receive the reward and distribute it depending on the invested assets percentage.
Fishermen can be thought of as light validators that oversee other validators. If a fisherman successfully reports a misbehaving validator, then that validator’s stake is slashed and the fisherman receives part of it as a reward. Fishermen will operate locally (shardchains) and although requiring less computing power than validators to operate, it will still be non-trivial.
Collators act as a support function, by suggesting candidate blocks for inclusion in the blockchain to validators - one can imagine them as the validators’ associates. Collators are paid by the validators for the services they provide - via sharing a % from the block rewards. Collators will only start appearing as the ecosystem grows.
Actor Required node Economic incentive Participation
Validators Full node Validation reward Validate blocks
Fisherman Full node Partial payout of slashed stake Masterchain tx to publish invalidity proofs
Nominator undefined Partial payout of validation reward Provide validator with capital
Collator undefined Partial payout of validation reward Point out shardchain block candidates
Figure 25: Classification of integral consensus parties by defining characteristics.
2.1.1.8. Programming on TON; Fift
TON is proposing a new programming language to run at its core; Fift. While technically multiple types of VMs and programming languages can be deployed on different
77 We view this design decision favourably, as Polkadot’s design although not yet in production, is theoretically the most well balanced design proposed in PoS thus far.
31
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“Fift was chosen workchains, Fift is designed specifically for developing and managing TON smart contracts on workchain_id=0 (the masterchain). Fift is an abbreviation of Forth - a 50 year for high usability old programming language, known for simplicity, extensibility, and enabling powerful in terms of programs. In this case it appears it was chosen for high usability in terms of nоdes’ nоdes’ cоnnectiоns - acting as the nuts and bolts that bring the “everything is a bag of cells” design philosophy to life. To illustrate, a develоper can take twо different nоdes оr even cоnnectiоns - part оf nоdes and cоnnect all the necessary parts in their smart contract using Fift, and acting as the extending that logic any part оf the prоgram can be divided and assembled backwards. nuts and bolts Perhaps an assertion one can make - besides the functional benefits of choosing Fift as that bring the the core programming language for TON - is that the core developers want to (i) increase ‘everything is a their level of control over the core protocol and/or (ii) create a mechanism for self- bag of cells’ selection, such that only the worthy become part of the “small council” that decides the fate of the core protocol. To illustrate, Fift employs Polish notation, which at least on the design surface does not have any functional benefits. In fact, if anything it makes learning (let philosophy to alone mastering) it, all the more challenging. This increases the potential risk for protocol life.” level mishaps, as it is both easier to make mistakes - even as an experienced developer - and harder for third parties to audit the contracts deployed. Given that blockchains have value transfer in their core, both the risk and the EV of critical failures, increase.
2.2. Best of the rest; other key pieces of the TON stack
The information that TON has released at the time of writing also foresees the following on- and off-chain applications that complete the vision for the network stack. To avoid
overload, this section will be limited to brief descriptions of these:
2.2.1. TON Payments
“Workchains A highly secure, fast and decentralized payment system with a transaction speed will be able to comparable to that of Visa and MasterCard. Payments are generalizable to any unit of account - i.e. it will facilitate the transfer of far more than just Grams. Although it support both introduces innovations such as smart payment channels that enforce smart contract logic free and paid on state channels, and an optimization for the number of nodes necessary to securely support such as system, it still remains at the concept stage and will not be launched state storage.” until after the network is up and running.
2.2.2. TON Storage
File-storage technology available for storing files with torrent-like access technology and smart contracts used to enforce availability. TON Storage employs ‘Merkle trees’ to store hashes of the data, allоwing data operations at high speeds. The module is also the core data store technology for the TON Blockchain. In its simplest form, it allows users to store files off-chain, by keeping on-chain only a hash of the file to be stored,
32
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
and possibly a smart contract where some other parties agree to keep the file in question “Validators are for a given period of time for a pre-negotiated fee. Workchains will be able to have both free and paid state storage. Paid (like Ethereum has) will make users to pay for every also expected to action - transferring coins, for instance. Free portions are going to be limited by the act as threshold set by workchain logic - i.e. one might be able to run small transactions without ‘caretakers’, fee on some workchains. Validators are also expected to act as “caretakers”, tending to removing any accounts or smart contracts that are not utilized, to decrease the tending to blockchain size. removing any accounts or 2.2.3. TON Networking smart contracts TON’s networking architecture is built on the Abstract Datagram Network Layer (ADNL), that are not a layer that regulates how all the TON nodes connect and share data. ANDL itself is an utilized, to abstraction over the commonly used UDP protocol. Further up the networking stack, one finds a Reliable Large Datagram Protocol (RLDP), along with a series of Gossip and decrease the Multicast protocols. The former is a transport layer abstraction that optimizes for speed, blockchain size.” while the latter moderate communication between use case specific clusters in the greater network (e.g. payment channel networks, storage networks, block propagation networks etc). The networking stack is completed by a Kademlia-like distributed hash table (TON DHT), that every TON supporting node across the core protocol or services and apps on top of it, will be using78. What the table effectively does, is to keep reference of every node in the network, so that queries are fast and effective.
2.2.4. TON DNS
A service to assign human-readable names to accounts that will allow users to access “TON DHT decentralized services as easily as browsing the World Wide Web. TON’s DNS will allow all the TON clients to quickly find and connect to the Messenger service, increasing the helps keep security and overall privacy of the Messenger - which also potentially explains the reference of decisions to design a DNS from the ground-up. every node in
2.2.5. TON Proxy the network, so that queries are 79 An anonymizer layer, similar to TOR . This layer can be used to create decentralized VPN fast and services and blockchain-based TOR alternatives to achieve anonymity and protect online privacy. In place to primarily protect large entities on the network (e.g validators) from effective.” DDoS attacks.
78 The DHT can be used as a “torrent tracker” for TON Storage, as an “input tunnel locator” for TON Proxy, and as a service locator for TON Services. 79 Tor is free and open-source software for enabling anonymous communication. The name is derived from an acronym for the original software project name "The Onion Router".
33
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
2.2.6. TON Services
“TON Services This is a pluggable application/utilities layer (akin to AWS’s or Azure’s suite of cloud is a pluggable services) that will initially is expected to be seeded with utility by TON, with modules such as DNS, Storage, Proxy and Payments discussed above. These services will live in a utilities layer, spectrum between purely on-chain and purely off-chain based and will be connected with akin to AWS’s one another and the TON Blockchain, through the TON Networking layer. The purpose of or Azure’s suite this layer will be to provide valuable tools for developers, while not requiring a direct interaction with the base layer (i.e. the TON Blockchain). of cloud services.” 2.3. Developer communities & recent updates
2.3.1. TON core, TON Labs and the early settlers
“TON has two TON has two main research groups working on TON’s development; TON core and TON main research Labs. The two co-exist like Parity and Geth on Ethereum, working on the client in a language of their choice - TON core opting for C++ and TON Labs opting for Rust. groups working on its TON Labs have consistently publicly claimed their arm's length relationship with Telegram development; and complementary (to TON) nature of their solutions.
TON core and The business model of TON Labs follows the standard open-source model, and is based TON Labs. on providing network infrastructure and cloud and enterprise services. The two co-exist TON core, led by the Durov brothers, is using the main Github repository80 for delivering like Parity and code in stealth mode81. While stealth development has its advantages early on, members Geth on of the wider TON developer community we interviewed expressed frustration over the Ethereum.” unpredictability of it all. Further, we currently have no way of knowing exactly how many people are working with TON core at the moment, but assuming that each of the released
modules has a dedicated team, we can approximate that there are currently 10 to 20 core developers putting time towards protocol-level work.
What TON core has released so far: TON node and validator software a CLI based “lite-client”, with limited support for the Testnet a Testnet explorer
80 [Access here: https://github.com/ton-blockchain/ton] 81 No public development is allowed in this repo.
34
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
in depth technical documentation on the TON Virtual Machine, the TON “Developers in Blockchain, and Fift. the community TON Labs, on the other hand, are more open to the general public about future plans; at we have spoken the moment, they are reportedly working on extending the LLVM compiler they recently to, view TON released for TON, while later planning to develop two-way payment gateways for cross- blockchain interoperability. Labs as a high value-add What TON Labs has released so far in their SDL: entity.” a local node server and node wrapper an ArangeDB database queried via GraphQL compilers that take Solidity and C and partially C++ code and compiles it down to TVM bytecode82. In the near future, the LLVM compiler is also poised to support Rust - a C++ lookalike, modern systems-level programming language, focused on safe concurrency83, and SWIFT
Developers in the community we have spoken to, view TON Labs as the more approachable entity in this small universe, and have found all their releases to be high
value-add. Indicatively, the node wrapper they recently released, brings the sync time for nodes to less than a minute – down from ~10 hours, while the Arange DB database makes it so that Dapp developers will no need to interact with the blockchain layer of TON. TON Labs also announced a partnership with Wirecard84, back in April 2019, through which
they aim to develop a suite of financial services that link the legacy consumer finance world with TON.
Besides the two aforementioned entities, there are a few open working groups consisting “At the moment, of independent researchers and developers that are the early consumers of TON TON is live in infrastructure. Our research points to the largest of those communities being Copper- Bits85, while we also discovered an organization called Formony, that has published a Testnet mode, python API client for the TON blockchain. with approximately
2.3.2. Current state of development 100 nodes running.” At the moment, TON is live in Testnet mode, with approximately 100 nodes running - presumably many of which are controlled by TON core and TON Labs. Engineers in the validator community outside the two main hubs, reference the need for a lot of backward induction in order to set up a validator node in Testnet and an overall lack of streamlined
82 The compiler is still at 20%-30% completion rate, deeming what is out there as a bare-bones PoC. 83 In computer science, concurrency is the ability of different parts or units of a program, algorithm, or problem to be executed out-of-order or in partial order, without affecting the final outcome. 84 Wirecard is one of the world's leading full-service providers of products and services for electronic payments, servicing some of the largest fintech companies, including Revolut and Atom Bank. 85 [Access here: https://github.com/copperbits/TON]
35
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
communication channels, which make the process difficult. On the flipside though, we are informed that more recently TON core is proceeding with frequent update releases in the TON repo, in an organized and consistent fashion – a process well received by the early developer community.
“More recently TON core is pushing updates in the Github repo, in an organized and consistent
fashion.”
Figure 26: Snapshot of TON’s main Github repository and recent code releases there.
Telegram also recently announced the launch of a hackathon, with a total prize pool of 86 ~$400k . The competition is focused on (i) building smart contracts for the TON Blockchain, (ii) suggesting improvements for the TON VM, and (iii) finding and fixing issues on the Testnet. Participants are expected to implement at least one of the five suggested smart contracts, including multi-signature wallet, two types of simple TON DNS Resolver smart contracts, a synchronous two-party payment channel, and an asynchronous two-party payment channel. The response from the developer community has been overwhelmingly positive. As it stands, there are 64k members in the Telegram Contests group, while the announcement for the competition over 350k views.
“Telegram also announced the launch of a TON focused hackathon, with a total prize pool of ~$400k.”
Figure 27: Snapshot of the original announcement of the TON Contest.
86 [Access here: https://t.me/contest/102]
36
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
2.3.3. Looking ahead
At the time of its fundraise, TON had set out the grand vision of having the platform fully rolled out by 2021, at which point it would transition from “Telegram Open Network” to “The Open Network” and the open source community would take the reigns. However, given the team’s delivery history thus far, the current market conditions, the technical complexity in bringing all the pieces together and attracting developers both in-house and into the broader community on TON, we believe that delivery in 2022 (at the earliest) is a more likely outcome.
Date Events and Milestones Actual delivery Delay Q2 2017 Start of the development of TON - Q1 2018 Launch of Telegram External Secure ID - Q2 2018 Launch of the Minimal Viable Test Network of TON Q3 2019 1+ years Q3 2018 Testing and security audits of TON Expected in Q4 2019 1+ years Q4 2018 Deployment of the stable version of TON Q4 2019 1 year Q4 2018 Launch of Telegram Wallet Q4 2019 1 year Q1 2019 Creation of TON-based economy in Telegram Expected in Q1 2020 1+ years Q2 2019 Launch of TON Services, Storage and Proxy Expected in Q2 2020 1+ years
Figure 28: Updated roadmap as initially featured in the TON Primer.
According to information that members of the TON developer community have shared with us, TON core’s immediate development plans include the development of standard techniques and algorithms such as airdrops, timelocks, escrows, oracles, smart contracts, DAOs, voting process structure, governance, etc.
Over the longer term, we expect the team to release privacy features for secure transactions - however, at this point in time it is almost impossible to predict how this feature will look like and when exactly it will be delivered.
37
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Part 3: Discussion
TON has undeniable promise as a 5th generation blockchain designed to solve the key “TON has issues that those that preceded it uncovered in the blockchain space. At the same time, there are many challenging factors which will define the fate of TON and the Gram as an undeniable investable asset, including uncertainty in the macro environment, the project’s many promise as a 5th technical intricacies, the cryptoeconomic model that underpins the network’s native generation asset (the Gram), and the contempt that many governments have for the network’s front- end (the Telegram messenger) as a harbor for illicit activity. These issues are discussed blockchain.” in sections 3.1 to 3.3 below.
3.1. All those in favour...
3.1.1. A huge market opportunity
A little over 10 years ago, the Bitcoin whitepaper proposed a proof on solving the double “If delivered as spending problem, planting the seed for the first digitally native monetary unit. 5 years ago, Ethereum introduced the concept of programmability in digitally native, money-like intended, TON assets, opening up a blue ocean of opportunity with the idea of smart contract based can be the applications. As the space matured, the lack of a de-facto distribution platform - like the gateway for App Store is for the Web 2.0 world - has become an obvious hurdle to conquer. TON could be that platform. If delivered as intended, with 250M (and growing) active users, cryptoassets and the majority of whom are located in parts of the world where they cannot access the full the related breadth of financial services available in the West, TON can be the gateway for applications to cryptoassets and the related applications to “bank the unbanked”. ‘bank the The use case for a financial system enclosed in a messaging app is beyond any doubt, unbanked’.” courtesy of WeChat’s spectacular growth in China since 2011. It is estimated that in 2018, WeChat processed transactions north of $20T in value87 amongst an estimated 1B MAU88 - a tremendous feat in creating and capturing value in the short span of 8 years. Besides payments, users of WeChat can access a series of services that span 89 from managing credit cards to purchasing insurance . However, the WeChat model is not easily generalizable at a global scale, as the tall structures that govern Chinese society, allow for privacy violations that are not tolerated in other parts of the world.
This is where the combination of a net of interoperable, value storing blockchains, and the Telegram front-end can really shine, as a permissionless system that connects
87 [Access here: http://bit.ly/2n7o3Tj] 88 [Access here: http://bit.ly/2nSuxW8] 89 Other use cases include scheduling doctor appointments, paying traffic fines, bike sharing, booking transportation, checking crowd density, searching for online trends, making contributions, business communication and tracking, calling landlines or other mobile devices.
38
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“Assuming multiple blockchains, protocols and the apps and services built on top of these, with millions of users, where the only blockage to value flow is in the bridge between the legacy volumes similar fiat world and the cryptoasset world. to those 90 WeChat The growth of the DeFi ecosystem within Ethereum is a strong testament to the traction that the application side of the platform is getting. WeChat is proof that the user need processed in exists. All that seems to be missing is a platform to link the two. TON has the potential for 2018 and a 0.1% market leadership in this space, and Grams as the native protocol token have the potential transaction fee, to capture some of that value creation.
the annual value To illustrate, given the above figures, if one assumes that Grams will only be used to stream to secure the network and act as a vehicle for validators to capture ledger fees, at transaction 91 validators volumes similar to those of WeChat in 2018 and assuming a 0.1% transaction fee , the potential annual value stream is ~$20B. That said, while the promise is undeniably there, stands at there are many obstacles in the way of TON achieving this scale, which we explore in ~$20B.” section 3.2 of this paper.
3.1.2. A collection of promising technologies
Both at the protocol and application layers, TON proposes a collection of some of the best tech that has been iterated upon over the last 10 years in the blockchain world . While there is much to be proven still in the deployment of it all, we cannot help but ascribe merit where it is due. TON, along with NEAR protocol, will be the first sharded blockchains network to go into production. By splitting activity centers in the network into shards that are responsible for most of the local resource management, TON promises transaction speed and resource efficiency never before seen in a global distributed network of value - a sorely needed addition to the space.
At the same time, contrary to DPoS blockchains that centralize consensus among a closed group of validators and create opportunities for collusion, eroding the integrity of the chain92, TON’s architecture has budgeted both for performance and sufficient decentralization. Some of the more granular highlights from the architecture include (i) the Infinite Sharding Paradigm, (ii) the dynamic messaging system for cross-shard communication, and (iii) the efficient resource management via sharding that makes DevOps more manageable at the local level.
90 DeFi is an abbreviation of Decentralized Finance, and refers to a collection of on-chain applications that propose novel ways to translate banking and finance services, on chain. At the time of writing the total value of assets locked in DeFi, currently stands at 3M ETH (or ~500M USD), having increased by 450x in USD terms since October 2017. 91 For reference, Visa currently charges merchants 1.51% plus $0.10 for a swiped consumer credit card, while Stripe charges 2.9% and $0.30 for online purchases. 92 See ‘Rampant Collusion in EOS Exposed by Huobi Leak’ on Trustnodes [Access here: http://bit.ly/2oOEV1v]
39
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
3.1.3. A world class team on a mission
TON is brought to life by one of the most talented and resilient teams in the technology “The team has industry. Their track record speaks for itself; from releasing VKontakte and growing it to proven beyond 500M accounts93, to self-funding the development of the Telegram Messenger, to more than 250M active users with minimal marketing, and from championing the right of any doubt that Internet users to privacy, to proving the hardness of their encryption technology and they can build distributed server infrastructure during the Russian and Iranian ban campaigns in 2018, distributed the team has proven beyond any doubt that they can build distributed systems at a global scale. systems at a global scale.”
3.2. And those that stand undecided…
3.2.1. Many unknowns in the token economy
When abstracting the protocol layer complexity away from blockchains, what remains is “Bad token a predominantly ‘social’ technology. Blockchains are the vehicle for decades worth of economics can insights from the fields of Game Theory and Mechanism Design to be deployed in industry, and native protocol assets (tokens) are the glue of it all. Good token economics break an and token economy management can help a network flourish. Conversely, bad token otherwise economics can break an otherwise fundamentally sound network by incentivizing (or not fundamentally disincentivizing) adverse behaviours among the many stakeholders. In TON, while there are many commendable elements in the way the token economy is structured, there are sound network also - still - a lot of unknowns which could sway the pendulum towards either side for the by incentivizing network. adverse behaviours.”
Figure 29: Zaki Manian’s (Head of Research for TON’s competitor - Cosmos) take on social scalability in blockchains today.
93 VK is ranked 19th in Alexa's global Top 500 sites.
40
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“ The TON The aspects of the token economy that we think TON has gotten right are (i) the multiple use cases they have instilled in Grams and the circular token economy that it has Foundation will designed around them, and (ii) the dual token market (primary and secondary) that neither be programmatically controls for wild price fluctuations by creating clear arbitrage involved as opportunities between the minting facility and the open market. Given that a lot of the utility value in Grams is poised to come from payments, low volatility and an upwards validators, nor trajectory in value are very important over the long term. While businesses and enterprise will they be worldwide increasingly – albeit slowly - warm to the idea of transacting in tokens, extreme buying back volatility is a not-so-silent killer.
Grams.” On the other hand, there are some potential points of failure in the overall design that center around the poor initial distribution of Grams among the broader ecosystem. While we can only speculate about the propensity of early investors to also be validators, the relative concentration of Gram holdings - at network launch - is high. And to make matters worse, as Telegram informed investors in a recent letter, in order to comply with SEC guidelines, the TON Foundation and Reserve will neither be involved as validators in the initial phases of the launch, nor will they be buying back Grams, as the whitepaper initially suggested. This puts the network launch firmly between “a rock and a hard place”. The risk of Grams being classified a security is nullified at the expense of a strong network
launch.
“This type of With the Foundation out of the picture, and given the initial validator set is projected to centralization be heavily concentrated among those early investors that are adequately prepared, it is not a far cry to expect a very concentrated launch phase. With the minimum stake can deter new requirement at 100k Grams, and the relative sophistication required to run a validator entrants by node, the pool of candidates becomes small. This also implies that this limited pool will raising be accruing the early validator rewards, until the network attracts more supply-side participants94. At the time of writing, there are a little more than a dozen entities setting questions about up validator nodes in Testnet. This type of centralization can deter new entrants by raising the network’s questions about the network’s governance and its overall integrity. And if not diffused integrity.” quickly, can put the long term prospects of TON and the Gram as an investable asset in a precarious position.
3.3. And all those against…
3.3.1. A vision too grand
While an impressive proposition as a whole, the feasibility of TON is still a theoretical exercise. The team has been secretive in their endeavors since March 2018 and remains so, even as the network is less than a month away from launching.
94 For an illustration of the problems that arise from a centralization of a staking token economy, see a report on the WAX Mainnet token economics we drew up, in May 2019. [Access here: http://bit.ly/2nnVuBb]
41
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
The lack of robust communication channels combined with the high technical complexity, leave us wondering whether the vision of TON, as it has been proposed, is actually possible in its entirety. While we have no doubts about the talent of the development team, we do have concerns about the amount of over-engineering that seems to be going into TON. Below we present some of the aspects of TON’s architecture that could transform from big question marks to Achilles heels;
a) The Infinite Sharding Paradigm; as mentioned in Section 3.1.2, conceptually, the “While we have idea is nothing short of brilliant. Given, however, its novelty and the lack of a formal no doubts about proof and/or exposure to the code that translates it into a system95, we cannot help but question its feasibility. Over-promising and under-delivering is, sadly, almost the the talent of the norm in the cryptoasset space. For context, Ethereum started its journey as the development scalable world computer, only to find that they would have to build a new platform team, we do from the ground-up, a short 3 years after its initial launch, in order to fulfil that vision. That said, members of the validator community we interviewed, noted that early have concerns indications in TON’s Testnet, are that the principles of the Infinite Sharding Paradigm about the seem to work, and expressed their excitement to see the full range of possibilities it amount of over- might enable. engineering that b) TON Networking; it is questionable whether the network can hit the throughput and seems to be latency targets it has set in the whitepaper, while a five second validation interval for going into block creation may not be - at all - necessary. While theoretically possible, with nodes TON.” placed thousands of miles apart, synched on a customized networking stack, such short intervals could lead to poor network performance (e.g. missed blocks)96. Again - theoretically, according to the Infinite Sharding Paradigm, these blocks would heal without the need for a rollback, so there would be no collateral damage. The feasibility of it all though, is still a question mark.
c) Governance; getting governance right is crucial in decentralized open source networks. A poorly governed system will fail to attract developers and users necessary to sustain itself over the long term. So far, little is known about how governance will be organized on TON. Given the similarities the consensus design in TON has with Polkadot, and the scale the network hopes to achieve, it is likely that coin voting will have a part to play in TON governance97. In that case, a TON Foundation actively diffusing the concentration of holdings and getting multiple parties involved to preserve the integrity of the network would be positive. The large initial concentration of holdings along with a feature of TON that allows users to open direct (and possibly private) payment channels with validators - deeming bribes a
95 As of yet, we have come across no information that points to the TVM supporting the cellular structures that power the Infinite Sharding Paradigm. 96 We understand that a custom networking stack is necessary to improve the hardness of the network, and given the team’s experience keeping a network under attack from multiple governments (often concurrently) live, we believe that there are good reasons behind those particular design decisions 97 We are firm in our belief that rough consensus [see here: https://en.wikipedia.org/wiki/Rough_consensus] doesn’t work for large scale systems as well as formal on-chain governance.
42
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
possibility, are potentially serious confounding factors that could compromise the
network’s path to maturity.
“The fact that d) TON Storage; the module lives further up the stack from TON Networking and through all blockchain its dependency is subject to all the adverse effects that the unknowns in the networking layer could cause. Further, TON Storage is a module to be developed further down the history - except line and given the delay in the delivery of TON, an early version of it is unlikely to go live for the most before Q4 2020. Furthermore, it is possible that in the initial phases of the rollout, all recent 2 months Storage nodes would be run by the Foundation. The fact that all blockchain history - except for the most recent 2 months - will be deposited in TON Storage, is a single point - will be of failure. While we understand that in the spirit of homogeneity, certain pieces of the deposited in TON puzzle need to be built from the ground up for the whole to work, we fail to find a TON Storage, good justification for this particular design decision. Provided that distributed cloud storage networks have been in iterative production in the blockchain space since at least is a single point 2014, we believe a better design decision to be an integration with one of those (e.g. the of failure.” Tardigrade Cloud from Storj)98. This type of decision would enable TON to go live with distributed storage ready to go, and increase the overall decentralization in the network.
e) Security & Privacy: As an encrypted messaging service, Telegram has not escaped public criticism about the encryption techniques they have used over the years. Further, while their track record of protecting their users’ data is good, it is not without stains. TON employs SHA-256 and SHA-512 as encryption hashes for the Blockchain and
Passport respectively. Both are known to be susceptible to brute force attacks, and while pluggable and open to discussion, this particular design decision looks questionable. It is likely that the overall security of the system will be hardened with TON Proxy going live - a building block estimated to be released sometime in 2020/2021. Moreover, engineers in the developer community we spoke to, mentioned that deploying SNARKs and Bulletproofs on TON contracts will be a relatively simple process.
“Telegram has not escaped public criticism about the encryption techniques they have used over the years.” Figure 30: Edward Snowden’s tweets echoing the overall public sentiment against Big Tech’s mishaps, relating to sensitive consumer data.
98 [Access here: https://storj.io/]. Without having an insight into the deep technical details of the project, we believe that modifying TON Proxy and TON DHT to be compatible with an existing distributed storage network, would both harden the network and save TON precious time and resources
43
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
f) Execution: given all of the above, along with the limited size of the core teams and the “TON is already limited community outreach the TON Foundation has performed thus far, execution an estimated 1+ becomes a major risk for TON. While the team’s talent and experience are stellar, this might just be their most ambitious project yet, in a space that consists of several moving targets. years behind TON is already an estimated 1+ years behind schedule and it is hard to imagine how the schedule and it window from roadmap to delivery will become narrower. Protocol level talent is hard to find is hard to justify in the blockchain space and the best developers are vying to start their own projects. The small pool of the next best, join those projects. Community building and good external how the communication are differentiators in attracting that level of talent, and thus far it appears window from that TON has been underperforming in that aspect. Failing to attract developers capable roadmap to of executing the roadmap will thus add to the overall execution risk. delivery will narrow.” 3.3.2. An unstable environment for cryptoassets
At this stage, Grams are “risk assets”. The Mainnet launch is coming at a time where both the macroeconomic and the intra-industry environments seem to be hanging in the balance. After a strong H1 in 2019, the cryptoasset market capitalization has retracted by 42% from Q2 highs at the time of writing, while the yield curve remains inverted for the 3rd quarter in a row. While Bitcoin may be in the early stages of becoming adopted as a safe haven asset, the same cannot be said for anything else in crypto.
Concurrently, at the time of TON’s launch, it is possible that the SAFT 1, and possibly the “The SAFT 2 investors in Telegram will be in profit. Given the general macro uncertainty, many cryptoasset of these early investors are likely to at least want to retrieve their cost basis upon finding liquidity. Were that to occur, the potential supply of Grams (assuming an average price of market cap has $2 per Gram)99, points to 2/3 of the unlocked coins flooding the supply side of the market retracted by 42% surface, over the course of Year 1. At the assumed average price level, this translates to from Q2 highs, $2.3B worth of Grams hitting the market. For the price to appreciate beyond the assumed $2 threshold, the supply of Grams would need to be met by more than $2.3B in demand; while the yield which translates to every single Telegram MAU buying at least approximately $10 worth of curve remains Grams. If we assume that only a 10% of that MAU base will be interested in purchasing inverted for the Grams, each would have to contribute $100, and so on. 3rd quarter in a row.”
Figure 31: The 2 yield curve inversions since 1990 (excluding the current) have successfully predicted upcoming recessions.
99 This is a purely illustrative figure.
44
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“It is unlikely Considering that the cryptoasset market is currently at a crossroads, with Bitcoin alone having shed over $40B of its market cap over September 2019, the retail speculative that Grams will appetite for Grams might not materialize in the initial stages of the float. Moreover, find real utility considering the early stage of the project’s development, it is unlikely that Grams will find in the Telegram real utility in the Telegram Messenger ecosystem in Year 1. And the fact that over Year 1 Messenger the circulating supply is subjected to the largest increase in the asset’s life (approx. 2.5x increase from launch until the end of Y1), certainly won’t do the project any favours. ecosystem in Year 1.”
Figure 32: Projected increase in the circulating supply of Grams.
3.3.3. A developer unfriendly experience
“Members of the Developer talent in the blockchain space is scarce and community building in Layer 1 validator protocols is often as important as protocol level work. TON seems to - thus far - put community involvement on the backburner, frontloading the platform build. There are few community communication channels with the TON core developer team, and those that are available describe the are of low quality. Members of the validator community that we have spoken to - process of unaffiliated with TON core or TON Labs - describe the process of setting up a validator in the TON Testnet as “requiring a lot of reverse engineering” as “the documentation is setting up a sparse and difficult to consume”100. Their evaluation is that without experienced validator node in developers driving the project, they would not be able to get as far as they have gotten Testnet, as and expressed concerns that due to the overall difficulty it is likely that the network will launch with a highly centralized validator set. requiring a lot of reverse Moreover, developers interested in TON are faced with the problem of ‘abnormal engineering.” programming’. The complexity of writing and testing a code in Fift is already proving challenging, according to reports from members of the early developer community. Until the compilers are ready for prime time101, Fift will be the only option to write smart contracts on TON, reducing the pool of available talent to those interested in low-level programming. To provide some colour here, TON Labs engineers we interviewed,
100 Conversely, engineers from TON Labs, describe an easier experience of setting up validator nodes in Testnet. 101 As we understand, the Solidity compiler exists only in a very limited version for now - i.e. it isn’t fully supported, while no decorator or modifier is present. Everything is working to the extent of 20-30% of the full functionality.
45
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
referenced that equivalent processes take 10x longer in Fift, than they do when “Equivalent programming on Ethereum with Solidity. That said, they also highlighted that Fift is an elegant and highly efficient language, and underlined their conviction that over the long operations, term, this is will pan out to be an excellent design choice for the network. currently take 10x longer in Be that as it may, TON currently remains without a debugging environment, which in combination with the relative difficulty of programming in Fift, make it near-impossible to Fift, than they write any meaningful programs. Further, even in the presence of a fully functional Solidity do when compiler, porting smart contracts from Ethereum to TON, will be a non-trivial process, as programming a relative redesign will be necessary in order to adapt the contract’s rules to TON’s logic – the same will most likely be true for smart contracts transitioning to Ethereum 2.0. on Ethereum with Solidity.” In addition, as discussed earlier in the report, independent third-party audits of smart contracts will likely be very expensive, as only a few highly specialized parties will be able to carry out that type of work. Tezos illustrates the problems that a developer unfriendly programming language can introduce - the smart contact platform leverages Michelson, a low level programming language unique to Tezos. The result? A dearth of smart contract applications built on Tezos102. Arguably, Layer 1 chains that leverage WebAssembly at the Virtual Machine level and more widely adopted programming languages such as Golang, sacrifice specificity (and to some extent performance) for generalizability; however, at this early stage of development in the blockchain ecosystem, generalizability might be necessary in order to attract talent and create a developer friendly experience.
Figure 33: A section from the lite client documentation.
Developers tinkering with deploying smart contracts in the TON Testnet have also “Reading from reportedly been facing issues with the fee structure of the TVM. Reading and writing to the TVM is expensive, as every deployment requires a fee to be submitted to validators. and writing to This not only is a discouraging factor for less well capitalized entities or experimenters, the TVM but also makes functions that require a high volume of interactions (e.g. on-chain order- appears book storage on a DEX application) unviable. Compounding on that point is TON’s commission-fee system, which at least at the moment makes it hard for developers on expensive.” TON to calculate the cost basis for their operations. As of yet, no rules have been set on
102 As of October 6th, 2019 Tezos had only has 108 contracts that contained code. This compares to over 11,000,000 contracts with code on Ethereum.
46
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
how the fee structure will pan out. All that is known is that the final fee will depend on the number of used nodes and the number of shards the users’ message has travelled through in order to get to its final destination103.
3.3.4. A series of competitors
“On the By merit of the broad spectrum TON covers across the technology stack, it faces messaging app competition in multiple different layers. TON’s primary competition is concentrated in the messaging application and the blockchain layers of the stack. On the messaging app front, front, Facebook’s Libra is arguably the most serious threat, followed by a series of Facebook’s smaller/regional players that are also integrating cryptoassets and blockchain wallets in Libra is their offerings (with Kakao/Klaytn in Korea and LINE in Japan being the most notable examples). arguably the most serious All of the aforementioned, while providing similar solutions, fail to translate the integrity threat.” and censorship resistance that truly decentralized systems bring to the table, by introducing a centralized front end. Telegram’s architecture is unique in that it makes no such compromise, while introducing a much broader value proposition than just
payments. Where we see Telegram somewhat lacking is in the volatility of Grams as a
native asset. All messenger layer competitors are introducing some fiat equivalent as the
rails for their payments feature to run on - arguably a more attractive proposition for the
user, both in that it provides assurances for value preservation, and in that the user is
required to make less effort to become familiar with the new feature104.
“Where we see That said, the benefits of building a system that allows for a consensus mechanism with Telegram wide participation from different parties are starting to show. Recently, 3 key members of the Libra Association105 (Visa, Mastercard and Paypal) are reportedly considering somewhat retracting their participation106 following the strong backlash the project received from lacking is in the US and European lawmakers, due to its effective singular control by Facebook. Besides volatility of its - provisionally - decentralized architecture, Telegram achieves sufficient differentiation, both drawing from the geographies where the core of its user base is Grams as a located, as well as the fabric of behavioural preferences that its user base exhibits (e.g. native asset.” privacy first, crypto native, censorship resistance).
Considering the competition on the blockchain layer of the stack, TON competes more closely with Ethereum (2.0), Cosmos and Polkadot along the interoperability and scalability vectors107. Each of the contenders have a unique differentiator about them; Ethereum has the state and the developer community, Cosmos has Tendermint and
103 With reference to the slow path routing of the IHMR. 104 The complexity that underpins cryptoassets is known to cause a very sharp dropoff in user engagement, very early in the funnel. 105 The consortium of enterprises that would run consensus on the Libra blockchain (similar to DPoS in EOS). 106 [Access here: on.ft.com/2OkoTaD] 107 For a feature for feature comparison between the 4 competitors, please refer to the Appendix.
47
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
multiple iterations behind it and Polkadot has the most mature and developer friendly “Ethereum has design, and a strong effort on community outreach108109. However, while in development for longer than TON, all the above are still far from maturity, with Ethereum 2.0’s Beacon DeFi and the chain110 expected in 2020, Cosmos’s IBC is expected in late 2020111 and Polkadot’s developer launch expected around the same time. That said, all three projects have engaged community, developer communities working on different implementations of the technologies, both on the protocol and the application layers. What they currently lack, is 250M active users Cosmos has to spoon-feed their technology to, through a familiar interface. We believe that if all the Tendermint, (significant) risks mentioned in sections 3.2 and 3.3 are sufficiently mitigated, TON is in Polkadot has a unique position to catch-up with and potentially outpace its competition. Substrate; TON has the 3.3.5. A regulatory mountain to climb users.”
Last, but certainly not least, is the regulatory minefield that most blockchain projects have to navigate. We estimate that approximately 60% of Telegrams MAU base is distributed 112 between Iran, Russia, Brazil and China - all markets where heavy government intervention in industry is commonplace. With concerns about the classification of the Gram as a security under SEC regulations recently being further diffused, we believe that most of the regulatory headwinds Grams will likely face will come from Telegram’s rocky history with the local governments.
Figure 34: Local currencies have performed poorly against the USD in at least the past 5 years, increasing the appetite for capital flight from the local economies.
108 Tendermint Core, is a high-performance, consistent, secure PBFT-like consensus engine, where strict fork- accountability guarantees hold over the behavior of malicious actors. Tendermint Core’s BFT consensus algorithm is well suited for scaling public proof-of-stake blockchains, and is already in use from various projects, including the Binance Chain and DEX. 109 Substrate is arguably the most mature SDK in the blockchain space. Developers can use Substrate’s modules to spin up parachains (the equivalent of workchains in TON) on the Polkadot ecosystem in minutes. 110 The Beacon chain is the equivalent masterchain in TON, and is the subject of the first out of four or five delivery phases that will be required to complete Ethereum 2.0. 111 The IBC (inter-blockchain communication protocol), is a protocol that will allow application-specific blockchains that are built using the Cosmos-SDK or Tendermint BFT to connect to the Cosmos Hub and interoperate with other blockchains that are connected to it. 112 [Access here: http://bit.ly/2ocHcUg]
48
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
“We believe that To make matters worse, Telegram has been targeted several times as a harbor for illicit most of the activity, including terrorism. A recent report by MEMRI, found that terrorist organizations like ISIS and Hamas are using the Messenger to run recruiting campaigns and raise regulatory donations from supporters in cryptocurrency. As mentioned earlier in this report, headwinds decentralizing the network as quickly as possible will help de-risk the project and separate it from Telegram, the entity, making Grams a harder and more censorship resistant TON will likely medium of exchange. face will come from Telegram’s Given Telegram’s history with the authorities in Iran and Russia specifically, we can assume that it will be hard for local governments to cut off access to the platform. rocky history However, further assuming that Iran, Russia, China and Brazil will be key geographies in with local driving adoption for TON Payments, and by extension the Gram, we estimate a relatively governments.” high risk of government intervention in cutting off the fiat on-ramps to the TON economy.
As laid out by the whitepaper, there will be two gateways to the Gram economy; the primary market, where presumably users can interact with the TON Reserve and purchase Grams via the Telegram app, and the secondary market, namely native crypto exchanges, OTC desks and P2P exchanges. While the secondary market can prove difficult to tame, the primary is easy to censor. The local currencies have shown continued weakness compared to the USD over the past half decade, and as such we expect both the demand for Grams and the incentive for government intervention, to be quite high113.
113 Indicatively, Iran recently outlawed trading Bitcoin and other non-sovereign money-like assets, while Localbitcoins - a popular P2P exchange for Bitcoin, had to halt operations in the region.
49
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Conclusion
The fundamentals of TON, the magnitude of its fundraise, the caliber of the team, and “Baking a the shroud of mystery that surrounds the project, have made the launch of the network the most highly anticipated launch of the last two years. Its fundamentals promise the philosophy of dawn of a new era in the world of blockchain facilitated networks of value. The user base openness early of the messenger, as an endowment, hints to a probability distribution more favourably on in the skewed towards a positive outcome for the Gram token economy114. development of Today, Ethereum works well for the developers that want to build high dollar value the network, transactions and low throughput applications, but leaves much to be desired in the without opposite domain. Platforms such as Cosmos, Polkadot and TON open up the possibility of low dollar value and high throughput applications, while being able to reference pre- compromising existing, valuable state from other networks through their interoperability features. Yet, on the core the road to the promised land appears thorny. tenets of privacy
The benefits of the Web 3.0 movement are found in its potential for enabling open and security, innovation. Early indications of this are found currently in Ethereum, where new smart will improve contracts are building on the state smart contracts deployed before them have accrued, TON’s to create a sprawling ecosystem of value flows, that spins up network momentum that, in turn, further attracts more talent and capital. competitive position.” Telegram appears - initially - as more of a walled garden. While introducing a crypto layer on top of the Messenger might increase the value of the present ecosystem by unlocking new features for Telegram’s users, it is hard to see how TON will transcend this narrow definition without opening up to the world.
Given the competition that TON faces in the blockchain/network layer, baking a philosophy of openness early on in the development of the network, without compromising on the core tenets of privacy and security, will drastically improve TON’s position in the competitive landscape - and pave a clear path for value creation and capture. Failure to do that, will likely lead to “tons” of wasted potential and will most likely mark one of the largest fundraising efforts in tech history, as little else than socializing the cost of running the Telegram Messenger.
Irrespective of outcome, TON’s launch, early-life and overall development will undoubtedly be full of learnings for entrepreneurs and investors alike. We, at Decentral Park, are looking forward to supporting the initiative as it finds its way in the world.
114 A similar experiment, the KIN token - courtesy of the Kik messenger, with an estimated 15M MAU, failed to get its token widely adopted by its user base. Telegram, with 15x the MAUs of Kik, has better chances of conjuring up the network effect necessary to sustain TON’s token economy. To provide some colour, Kik announced that it will proceed to shut its messenger down and downsize by 80% in order to continue pursuing its legal battle with the SEC, over whether its ICO was an unregistered security sale. [Access: http://bit.ly/2lkXBob]
50
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Appendix
Table 1: Analysis of TON’s competition in the blockchain space, point by point.
TON Ethereum 2.0 Polkadot Cosmos
Website test.ton.org https://www.ethereum https://polkadot.network/ https://cosmos.network/ .org/
Year 2017, (2018) 2013, 2015 2016, (2019) 2017, ? announced and year deployed
Community Telegram - 3923 Telegram - 11562 Growth members115 members117
Twitter - 31300+ Twitter - 28300+ followers116 followers118
Type Decentralized Decentralized Decentralized Decentralized
Consensus PoS BFT PoW PoS BFT PoS BFT Algorithm
Support for yes yes yes yes Arbitrary Code
Blockchain multiple single multiple multiple System mix heterogeneous heterogeneous
Interactions tight loose loose Between Blockchains
Speed 50 000 - 70 000 2000 - 4000 Custom User-defined 100 000+ transactions Transaction transactions per second transactions per protocol per second with the help (Visa and MasterCard second of Tendermint Protocol speed level)
Block Creation 5 seconds 12,5-15 seconds Custom User-defined Speed protocol
Block Delivery Delivery based on 1-2 15 seconds Custom User-defined blocks protocol
Validators App. 100 validators will New validators are At the early stage, the be selected monthly for a selected every 24 hours. Cosmos network will period of 30 days by the Validator can be any have 100 validators (who system among all person, after registration has the largest stake in participants who have and compliance with Atom), but every year the sent contributions to the number of validators will
115 Polkadot Telegram Chanel [Access here: https://t.me/PolkadotOfficial] 116 Polkadot Official Twitter Page [Access here: https://twitter.com/polkadotnetwork] 117 Cosmos Telegram Chanel [Access here: https://t.me/cosmosproject] 118 Cosmos Official Twitter Page [Access here: https://twitter.com/cosmos]
51
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
smart contract. conditions.120 grow until it increases In order to become a 300. Any member of the voting validator, the user network can offer his needs at least 100K candidacy as a validator. GRAMs. To participate in the validators election round, the user needs first to lock (stake) the tokens. Funds can be released only after the start of the next round. Thus, participation in each new round is, firstly, not automatic, and secondly, the user cannot participate in consecutive rounds with the same funds.119
Cross- Two types: Through relay chain Cosmos -> Cosmos Hub Shards 1 - Long way -> Tendermint Messages 2 - Fast Way Routing Hubs Model
Errors in Shards Errors are managed by Polkadot will return all Everything depends on Patch use chains up to the point the choice of validator when the invalid transaction was made.
Developer fift + TONVM - low level Works: ● SDK: Tools language ● Rust-based DSL Tendermint team (templated) develops the number TON Labs: ● + special Rust of basic modules, Style Guide from which are required for Solidity flavor compiling Polkadot - Cosmos Hub. to (based on LLVM) https://wiki.polkadot.net ● Cross zone Vanilla C language work/en/latest/polkadot/ assets transfer - not support build/rust-style-guide/ working yet https://github.com//TON Copper-Bits
119 TON Blockchain Test Network — files and resources - https://test.ton.org/download.html 120 Official Polkadot Community - https://polkadot.network/ru/community/
52
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Figure 1: ATON’s methodology on arriving at a fair value for Grams
Figure 2: Illustration of the various forms of value flow in modern day blockchains.
53
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Figure 3: Graphical illustration of a hypercube structure.
Figure 4: Graphical illustration of the Sawtooth PBFT consensus algorithm.
Figure 5: Comparison of message complexities of some classical BFT consensus algorithms.
54
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Figure 5: Estimated communication cost in megabytes (per node) for varying batch sizes in HoneyBadgerBFT.
Figure 6: Throughput (transactions committed per second) vs number of transactions proposed in HoneyBadgerBFT. Error bars indicate 95% confidence intervals.
55
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Figure 7: Latency vs throughput for experiments running HoneyBadgerBFT over Tor.
Figure 8: A high-level view of PBFT, Q/U, and Zyzzyva.
56
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021
Author:Author: Elias Elias Simos, Simos, Senior Senior Research Research Analyst Analyst at decentralpark.ioat decentralpark.io
Important Disclosures
Decentral Park Advisors, LLC (“Decentral Park”) has prepared this material for informational purposes only. This material does not constitute trading strategy or investment advice, or an offer to buy or sell, or a solicitation of any offer to buy or sell, any security or other financial instrument.
The information and opinions contained in this material are those of Decentral Park and are subject to change. We make no representation or warranty with respect to the accuracy, timeliness or completeness of this material and have no obligation to update it. We make no guarantee that any forecasts or forward- looking information will happen. The performance of any specific investment, sector or market included in this material does not reflect the expenses, fees and taxes generally paid with the active management of an actual portfolio.
Decentral Park, its affiliates or their employees or members of their families, may at times have a long or short position in the securities or financial instruments discussed in this material and may make purchases or sales of these securities or financial instruments while this material is in circulation.
Decentral Park does not provide legal or tax advice. Decentral Park is not acting as a fiduciary under either the Employee Retirement Income Security Act of 1974, as amended, or under Section 4975 of the Internal Revenue Code of 1986, as amended.
Investing involves risk, including the possible loss of principal. Investments discussed in this material are not be suitable for all investors. Alternative investments often are speculative and include a high degree of risk. Investors could lose all or a substantial amount of their investment. Alternative investments may be suitable only for eligible, long-term investors who are willing to forgo liquidity and put capital at risk for an indefinite period of time. Alternative investments may be highly illiquid and can engage speculative practices that may increase the volatility and risk of loss. An investor should create a customized investment plan, and any investment plan should be subject to periodic review for changes in individual investor circumstances. Past performance is no guarantee of future results.
This material, or any portion thereof, may not be reproduced, sold or redistributed without our consent.
57
Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021 Downloaded from www.hvst.com by IP address 192.168.160.10 on 09/30/2021